Gallup Analytics Data Session

CBA Information Technology Services Author

03/22/2017 Added

26 Plays

Description

Introduced by David Hartline, Elizabeth Steele, Gallup Regional Director for Northern America, will presents an information session about how faculty can use Gallup Analytics data in their research. As part of the generous donation by the Clifton Foundation and Gallup, CBA faculty have access to the Gallup Analytics platform, allowing access to nearly 100 years of U.S. data and 10 years of Gallup World Poll data.

Searchable Transcript

Search:

[00:00:10.808]Okay, everybody.
[00:00:11.901]Thank you for attending.
[00:00:14.216]I'm David Hartline of IT
[00:00:15.771]and Facilities Operation for the building,
[00:00:17.144]I'm very excited about what we're gonna learn about today.
[00:00:19.921]As part of the donation that Gallup made
[00:00:22.432]to the new building project, along with funding
[00:00:26.255]and other things, came access to this World Poll data
[00:00:30.292]and Analytics data, and we've been working with Gallup
[00:00:33.090]themselves, and with Elizabeth Steele,
[00:00:35.286]to try to understand more of what we have
[00:00:37.758]access to and how to use it.
[00:00:39.350]We have access to it now.
[00:00:40.683]You saw an email from Donny and one from me
[00:00:42.592]about how to get access to it,
[00:00:45.068]but now we can learn about how to use it,
[00:00:46.972]and so, thank you very much for attending,
[00:00:48.832]and welcome, Elizabeth.
[00:00:50.468]Thank you.
[00:00:51.301]Thank you so much, and thank you all
[00:00:52.656]for attending on a Friday afternoon
[00:00:54.412]before a Husker home game, go Big Red.
[00:00:57.008]I've got a special thank you to all of our
[00:00:58.142]technology partners who helped me set this up.
[00:01:00.516]It's never possible without people like you,
[00:01:02.584]so thank you for that.
[00:01:04.073]Our discussion toady will cover the data sources
[00:01:06.560]in Gallup Analytics, and also how to use the tool itself,
[00:01:09.764]so the first portion of the presentation
[00:01:12.364]will be a background on the methodology,
[00:01:14.184]so how the data is collected, how it's processed
[00:01:16.584]by Gallup, and how it comes to be in Gallup Analytics,
[00:01:19.268]and then, I'll do a really quick demo showing you
[00:01:21.312]som of the capabilities of Gallup Analytics.
[00:01:24.100]I will make sure to reserve 15 to 20 minutes at the end,
[00:01:27.808]but if you have any pressing questions,
[00:01:29.352]feel free to shout them out or raise your hand.
[00:01:31.420]I'm happy to answer them as we go along,
[00:01:33.596]and especially because everyone has such diverse
[00:01:35.558]research interests, we can more target it to you
[00:01:38.216]if you ask pointed questions, so feel free at any time.
[00:01:42.459]The Gallup Analytics tool is made up of three
[00:01:44.264]main data sources, so we have two data sources
[00:01:46.838]that are sort of our main data sources in the tool,
[00:01:49.256]and then we have another historical data set
[00:01:50.848]that you do have access to through Gallup Analytics.
[00:01:53.965]So the two main data sources in the tool
[00:01:56.224]are gonna be our Gallup World Poll
[00:01:57.849]and what we call our Gallup Daily Data.
[00:02:00.551]So, Gallup World Poll data is collected
[00:02:03.056]in usually about 140 countries every year.
[00:02:06.640]Depending on the penetration of phones, we do either
[00:02:10.016]face-to-face, or we do landline and cell phone interviewing.
[00:02:13.976]For example, in the US, we do a mix of about 50/50
[00:02:17.471]landline and cell phone interviews.
[00:02:19.642]Depending, we might change that to about 60/40
[00:02:22.838]when we're looking at sort of the overall composition
[00:02:26.156]of the sample we're getting back.
[00:02:27.893]For all of our face-to-face countries,
[00:02:29.859]we are stratifying to get a representative sample,
[00:02:33.339]which means when you look at the Egypt data,
[00:02:35.583]you're looking at the whole country of Egypt,
[00:02:37.775]you're not just looking at Cairo.
[00:02:39.690]So, in most countries, we're using a geographic variable,
[00:02:42.467]so a regional variable.
[00:02:43.863]I think the United States makes sense to talk about.
[00:02:45.783]So, we would use states as the regions,
[00:02:49.044]and then, we're using a population size to stratify as well,
[00:02:52.683]so we make sure we're getting enough urban respondence
[00:02:55.535]and enough rural respondence to truly represent
[00:02:58.013]the will of the country.
[00:03:00.211]We complete about 1,000 interviews every year.
[00:03:03.139]All of our interviews are trained the same way around
[00:03:05.731]the world, we use standardized interviewing techniques,
[00:03:08.687]if anyone is particularly interested,
[00:03:10.323]please come talk to me afterwards.
[00:03:12.102]We can talk about how the interviews are trained.
[00:03:14.489]But every interviewer around the world,
[00:03:16.927]whether they work directly for Gallup
[00:03:18.639]or they work for a vendor hired by us,
[00:03:20.454]does undergo that Gallup interviewer training,
[00:03:23.079]so we can be convinced that everyone is being administered
[00:03:26.591]the survey in the exact same way.
[00:03:29.487]One thing to ask, what do we do in dangerous
[00:03:32.486]and repressive countries?
[00:03:33.671]We do not survey in countries where we cannot
[00:03:35.286]get the permission from the government,
[00:03:37.357]or we believe the health and safety
[00:03:39.051]of the interviewers would be at risk.
[00:03:41.123]This year, we were not able to interview in Syria,
[00:03:43.626]but we carefully monitor all those situations,
[00:03:45.623]and we try to be representative of the world as a whole.
[00:03:49.217]The Gallup Daily Poll is gonna be a little bit different.
[00:03:52.387]So, the Gallup Daily Poll represents 1,000 completes
[00:03:54.891]of US respondents each and every night,
[00:03:57.099]so I was actually just at our Edgewood call center,
[00:04:01.223]and we have about 200 people there that are calling out
[00:04:04.267]every night to ask about people's well-being,
[00:04:07.163]their economic situation, and things like that.
[00:04:09.698]So, the Gallup Daily Poll is collected
[00:04:11.707]with that RDD, so random digit dial,
[00:04:13.941]which means we just purchase phone numbers
[00:04:16.191]from a sample vendor, and then we call
[00:04:18.314]until we have 1,000 completes.
[00:04:20.571]Something to know specifically about this tool
[00:04:23.109]is that all of the data in this tool
[00:04:26.011]has already gone through the cleaning
[00:04:27.967]and weighting process, so all the data in this tool
[00:04:30.323]is weighted to be representative.
[00:04:32.075]In the US, we use census targets,
[00:04:33.792]which is, I think, generally,
[00:04:36.451]what a lot of organizations are doing.
[00:04:39.439]In World Poll, we use what we call the best available
[00:04:41.779]information, so sometimes, that will be the
[00:04:44.075]central statistics office of a country,
[00:04:46.518]sometimes we'll use some DHS targets, and sometimes,
[00:04:50.315]we'll combine sources of information
[00:04:51.607]to figure out what those targets should truly be.
[00:04:54.703]Sometimes, governments have no idea on the dispersion
[00:04:57.455]of their population, so we do the best we can,
[00:05:00.035]and we do do some modeling, and to think about
[00:05:03.581]the true composition of the sample on the back end,
[00:05:08.848]so all this data comes out weighted.
[00:05:11.543]For the World Poll data as well,
[00:05:13.276]you can access both regional aggregates
[00:05:15.976]and global aggregates, and that data is weighted to
[00:05:18.757]what we call a population projection,
[00:05:21.193]so what that means is instead of doing a global average
[00:05:23.160]of all the countries, so, for example,
[00:05:25.919]China would get the exact same weight as Mauritius,
[00:05:30.457]we do a population adjustment, so we adjust
[00:05:33.342]for large populations and small populations,
[00:05:35.306]which means more populous countries get a bigger weight,
[00:05:38.126]and therefore, we have a true representation of the world
[00:05:41.214]as opposed to just an aggregate of countries,
[00:05:44.010]so all of those weights are being applied in the tool
[00:05:47.442]as soon as you run it, so if you are using something
[00:05:50.271]in Gallup Analytics, you can use it in a paper,
[00:05:54.430]you can use it in research, because you know
[00:05:56.121]that that number has already been quality-assured
[00:05:58.562]by us and that it's running using the proper weights,
[00:06:01.319]and the tool will do that automatically.
[00:06:02.883]You don't need to go and apply anything.
[00:06:05.772]The third source of data is gonna be
[00:06:07.071]what we call the Gallup Brain.
[00:06:08.975]The Gallup Brian is a composition of a variety of sources,
[00:06:11.691]so it's hard to talk about the specific methodology.
[00:06:14.729]When we do our own internal user analytics,
[00:06:17.391]we're finding that most people are using
[00:06:18.935]what we call our social series data.
[00:06:21.062]So, the social series data is collected
[00:06:22.775]in much the same way as our US daily data,
[00:06:25.249]so it's 1,000 completes, but we're doing a monthly
[00:06:27.439]topical study, so we call in March,
[00:06:30.351]and we ask people about issues of the environment,
[00:06:33.542]we call in February and we ask about race and minority
[00:06:36.671]relations in the United States, so those are more topical.
[00:06:40.012]We do have data going back to the 70s in the Brain,
[00:06:44.284]but it doesn't work the exact same way
[00:06:46.319]as those World Poll and G1K data sources
[00:06:49.555]just because the structure of the data
[00:06:51.688]is a little bit different.
[00:06:52.825]But I'll show you both of those.
[00:06:55.829]Before I continue, does anyone have any
[00:06:57.541]questions on the methodology?
[00:07:01.017]Not about the methodology,
[00:07:01.850]but how far back does the World Poll
[00:07:03.405]and the Daily Poll data go, if you wanna use one of those?
[00:07:06.315]So, the Gallup World Poll started in 2005 and 2006.
[00:07:09.660]We combined those years, and those are all in the tool.
[00:07:12.702]I would caution you that 2005 and 2006 have fewer
[00:07:15.916]countries than after that, and I actually don't recall.
[00:07:19.596]We'll look when we look at the G1K data.
[00:07:21.147]We have at least 10 years of G1K data,
[00:07:23.340]and it might be closer to 15,
[00:07:25.132]but we'll look at that when I pull up the G1K.
[00:07:30.202]And all of the data for World Poll is by the year,
[00:07:33.052]but G1K, it updates automatically every day,
[00:07:37.096]so you can look at a month, you can look at a day,
[00:07:39.802]you can look at a year.
[00:07:41.343]Unfortunately, the tool does not allow you
[00:07:43.130]to combine your own time periods yet.
[00:07:45.362]That's something that we're looking at for 2017.
[00:07:49.381]But I think with that, I'll start showing the tool.
[00:07:52.205]I know you have the email on how to access it.
[00:07:54.280]I actually, really excitingly, typed in
[00:07:56.538]analyticscampus.gallup.com, and it immediately
[00:07:59.226]popped up because you have the IP option,
[00:08:01.464]so if you're ever on a campus computer,
[00:08:04.310]you can just type in analyticscampus.gallup.com,
[00:08:07.046]and it will pop up.
[00:08:08.314]It will bring you to this main page,
[00:08:10.402]which allows you search in three major ways.
[00:08:12.742]I think they're pretty self-explanatory.
[00:08:15.196]I often start with geography, because I usually know
[00:08:18.264]that I am interested in a certain part of the world,
[00:08:21.750]and something to understand about the tool,
[00:08:24.071]it is segmented between the data sources,
[00:08:26.644]so you cannot run the US Daily data
[00:08:29.285]on top of the World Poll data,
[00:08:30.996]'cause they're actually kept separately in the system.
[00:08:33.214]You can certainly run them in the same user session,
[00:08:35.767]but you can't run them side-by-side.
[00:08:37.303]You do have to pick a source whenever you
[00:08:38.839]do any analysis, so I'm going to go to the World Poll data.
[00:08:45.814]My main job at Gallup is I am a methodologist
[00:08:48.004]on the World Poll, so I know that a little bit better.
[00:08:51.094]So, I'm gonna click worldwide.
[00:08:53.411]And you're going to look at the top,
[00:08:54.742]and I think probably these tabs at the top
[00:08:57.972]are pretty self-explanatory, and when we look at
[00:09:01.072]the analytics of how people are using this,
[00:09:03.321]they're spending the majority of the time
[00:09:04.754]on this Tables page.
[00:09:06.161]The Tables page is the only sort of
[00:09:07.792]possible way to cross-tab the data.
[00:09:10.369]The charts are also a visual way to analyze the data.
[00:09:13.747]We'll talk about these three features at the end,
[00:09:16.125]but I'm gonna go to the Tables page,
[00:09:18.541]and how it works is you click your data source,
[00:09:21.724]and at any time during your session,
[00:09:23.136]if you wanna change your data source,
[00:09:24.433]always look up here, click data source,
[00:09:27.287]and then you can add a geography.
[00:09:29.023]I'm going to actually add Armenia.
[00:09:33.453]And then, you're going to be able to look at,
[00:09:35.441]for World Poll, about 120 metrics,
[00:09:37.984]and they're gonna be in the drop-down,
[00:09:39.584]and so, you can kind of scroll down,
[00:09:41.423]and we're gonna look at, let's do the quality
[00:09:45.554]of healthcare in the city.
[00:09:46.898]And so, one thing that is important for any
[00:09:48.929]report or any project you're doing,
[00:09:51.180]if you want the exact question text,
[00:09:52.831]you can simply hover it, and it will tell you.
[00:09:55.430]We're looking at it formatted right now,
[00:09:57.618]and I'll discuss formatted versus unformatted later,
[00:10:00.690]but this will just give you the overall aggregates
[00:10:02.764]for worldwide, and then when you scroll down,
[00:10:04.961]we can see Armenia as well.
[00:10:07.440]So, those are, of course, applying the proper weights,
[00:10:09.438]and then you can look at the categories.
[00:10:11.715]So, if we look for worldwide, people who are satisfied
[00:10:14.915]with the quality of healthcare in their city,
[00:10:16.748]we're seeing about 58% that are satisfied,
[00:10:19.698]39% not satisfied.
[00:10:22.351]DKRF means don't know and refused.
[00:10:25.570]We combine those together.
[00:10:26.895]They are part of the overall total.
[00:10:29.633]We don't drop those responses, but that is something
[00:10:31.890]you can do, if you would like, in the Excel exports.
[00:10:36.818]So, we are looking at all the years of data for World Poll.
[00:10:40.146]I'm actually gonna clear this, and we're just gonna look at
[00:10:43.700]2015, so that's a little bit less to look at.
[00:10:48.967]And so, the analysis portion of the Gallup Analytics tool
[00:10:51.975]allows you to look at the data by the given demographics,
[00:10:56.408]so in order to do an analysis, you're going to add
[00:10:58.994]the demographics, and you can add as many as you want.
[00:11:02.021]I'm gonna add gender, just because it has two,
[00:11:03.602]and sometimes with limited room on the screen,
[00:11:05.926]it's easier to look at that way.
[00:11:07.621]So, you can choose to uncheck Aggregate
[00:11:09.765]if you do really truly wanna just look at the breakdown
[00:11:13.991]of gender, but this also shows you the aggregate, too,
[00:11:16.504]if you'd like to compare the groups to the overall.
[00:11:18.774]So, again, these are using all the proper weights,
[00:11:21.545]and you can look worldwide, you can look,
[00:11:24.726]I'm actually gonna add a region, so you can see that.
[00:11:27.842]So, I'm gonna add Eastern Africa,
[00:11:30.466]and then you can add the Eastern Africa aggregate.
[00:11:35.461]And then, that will add in as well.
[00:11:36.886]So, you can look at a variety of different geographies.
[00:11:39.637]If you want to share this with a coworker,
[00:11:41.960]with a student you're collaborating with,
[00:11:44.086]you can use this link button right here,
[00:11:46.135]and it will create a unique link that will
[00:11:48.105]bring you back to this table, so if anyone else
[00:11:50.551]is interested in seeing exactly what you saw.
[00:11:53.140]If you ever experienced a problem with Gallup Analytics,
[00:11:55.717]one great way to get, I think, slightly faster
[00:11:58.565]response times is if you provide the link to us.
[00:12:01.321]Often, we can help you understand, because we're
[00:12:03.958]looking at the exact same thing you're looking at.
[00:12:06.309]You can also choose to export this data,
[00:12:10.048]and it will come in a formatted Excel file,
[00:12:14.248]and I'm gonna show you the unformatted next,
[00:12:16.150]but if you want a formatted Excel file,
[00:12:18.417]perhaps you're gonna use it in your class
[00:12:21.492]to talk about changes in well-being around the world,
[00:12:24.434]that's a really quick way to do that.
[00:12:27.191]I'm gonna go back down.
[00:12:28.852]I'm actually gonna go back to the interface.
[00:12:31.149]What I would suggest if you're gonna do some analysis
[00:12:34.374]is to click this Unformatted button.
[00:12:36.008]So, the Unformatted button just means you're not
[00:12:38.042]gonna get any of that Excel formatting when you export
[00:12:40.855]the data, and you also are looking more at the raw numbers.
[00:12:45.087]It's really a lot of work.
[00:12:47.930]Everyone knows to format those percentages
[00:12:49.786]to get them right to actually use them.
[00:12:52.493]Perhaps you wanna build your own index,
[00:12:53.811]so I would definitely say any exports you do
[00:12:56.592]will be much easier if you click Unformatted to export.
[00:13:00.659]We're gonna go to, and if you're ever,
[00:13:02.651]if you wanna start over, as opposed to working
[00:13:04.419]from where you are, really quick, Reset,
[00:13:07.043]and you can start over.
[00:13:08.225]You can start over.
[00:13:09.569]So, I'm gonna look at the US Dailies
[00:13:12.786]to show you the chart features.
[00:13:15.290]So, the chart features will allow you to look at
[00:13:18.309]the data more visually, so we have a trend chart,
[00:13:21.872]and that will be basically looking at a trend over time,
[00:13:24.899]so I'm gonna add Nebraska, and we're gonna look at
[00:13:34.363]if it's becoming a better or worse place to live,
[00:13:36.679]and so, we're actually comparing Nebraska
[00:13:38.956]to the United States as a whole,
[00:13:40.801]so you can see the key here, we're green,
[00:13:43.396]this is the United States, and when you hover over any
[00:13:46.198]of the given points, it will give you the exact breakdown.
[00:13:49.300]One of the things you cannot toggle on these
[00:13:52.065]is you can't shrink any of the axes,
[00:13:54.890]so unfortunately, what you see is what you get.
[00:13:57.652]So, I apologize for that, and that's something
[00:13:59.236]we're also looking to fix in 2017.
[00:14:03.574]So, you can export these as well,
[00:14:04.843]so if you're making a Powerpoint for class
[00:14:07.246]and you want to make a point about, we have some questions
[00:14:11.376]about whether or not the economy
[00:14:13.125]is getting better or getting worse,
[00:14:13.962]this is a really quick way to create graphs.
[00:14:16.392]You can export it in a PDF form or a variety of
[00:14:21.632]image files, so that's gonna be the trend chart.
[00:14:27.101]We could add all the geographies
[00:14:28.930]really quickly if you wanted to do that.
[00:14:30.563]This'll be a little bit harder to see.
[00:14:33.002]But that's a really quick way, if you want to look at
[00:14:35.474]20 states, you can go by deleting them by adding them.
[00:14:39.803]But that makes sense, also, for a rank chart.
[00:14:43.674]So, the rank chart will allow you to look at any metric.
[00:14:46.145]It will allow you to look at the rank.
[00:14:48.558]I would say, as a researcher myself,
[00:14:51.768]there are some cultural differences
[00:14:53.849]in how people answer questions, especially around the world,
[00:14:56.934]so we find overwhelmingly that Latin Americans
[00:14:59.411]are much more positive than Eastern Europeans,
[00:15:02.662]even accounting for a variety of factors.
[00:15:04.555]They just more positively answer questions.
[00:15:06.834]So, I would caution you against one-to-one,
[00:15:09.057]building a rank chart and comparing.
[00:15:11.447]It's also good to think about,
[00:15:13.896]occasionally, we'll have some African countries
[00:15:15.695]that have experienced genocide popularly high on the list,
[00:15:18.577]and that is because their relative well-being,
[00:15:21.402]they consider very good.
[00:15:22.935]So it is a subjective evaluation of things like that,
[00:15:25.951]so keep that in mind before you
[00:15:27.900]maybe, perhaps make the claim that people
[00:15:30.458]in Rwanda are doing much better than people in Finland.
[00:15:35.457]Just few things to consider.
[00:15:37.441]So, that is gonna be the rank chart,
[00:15:39.689]and then we have the scatterplot.
[00:15:41.175]The scatterplot will allow you to look at two metrics
[00:15:44.241]at the same time, so I'm going to overlay with,
[00:15:49.690]let's do city satisfaction.
[00:15:53.272]Oh, okay, let's change this.
[00:15:56.644]We're gonna look at...
[00:16:03.970]Let's see.
[00:16:07.946](mumbles)
[00:16:10.975]And so, oh, we just clicked United States.
[00:16:12.739]So, that will enable you to look just
[00:16:15.410]at a scatterplot of multiple metrics.
[00:16:17.687]I actually think the World Poll data
[00:16:20.985]is a little bit easier to look at the scatter plots.
[00:16:25.714]We're gonna do, we're just gonna add...
[00:16:30.920]Freedom of Life, and then we're going to
[00:16:32.759]add one more, let's add...
[00:16:40.735]So, this will enable you to compare all the countries
[00:16:43.601]around the world on the two metrics,
[00:16:45.504]so we find that Uzbekistan is really high
[00:16:48.127]on freedom in your life and Economic Confidence.
[00:16:51.625]For a lot of the former Soviet Union countries,
[00:16:54.311]we suspect that there is a certain oppression effect,
[00:16:57.888]so when interviewers report to us that when they go
[00:17:01.641]to people's houses, there's a lot of questions on,
[00:17:03.722]"Is this from the government?"
[00:17:04.765]Like, "Where is this data going?"
[00:17:06.613]And that it could be that people are reporting
[00:17:08.971]higher numbers because they are concerned
[00:17:11.030]about the repercussions.
[00:17:12.977]We do work really hard to maintain the confidentiality
[00:17:16.361]of all our respondents, so all of the tracking sheet
[00:17:18.996]data on people's households, people's household size
[00:17:21.611]and their age, their identifying information,
[00:17:24.122]that's kept separately from the data,
[00:17:25.785]and we're also doing continual analyses to see,
[00:17:28.180]for example, if someone has a very high income
[00:17:31.465]in a very small geographic area.
[00:17:33.709]What we can truly report together that might be identifying,
[00:17:36.644]it might be potentially dangerous for an individual,
[00:17:38.947]so that's something we're continually
[00:17:40.913]monitoring and working on.
[00:17:43.439]We're gonna go to the map feature.
[00:17:46.433]Personally, internally, I used the map feature
[00:17:48.533]to populate the first piece of my Powerpoint slide,
[00:17:51.599]so if I am interested in explaining
[00:17:53.483]economic confidence around the world.
[00:17:55.859]We're gonna get a really quick heat map
[00:17:59.355]of economic confidence around the world.
[00:18:01.800]So, of course, the lighter colors mean a lower
[00:18:04.818]economic confidence, and then the dark colors mean
[00:18:07.403]a higher economic confidence, and then you can
[00:18:11.099]hover over it to see what the exact score is.
[00:18:14.201]I'm actually gonna do latest available.
[00:18:19.138]Sorry about that.
[00:18:19.971]I did not realize the cookie statement was up there.
[00:18:21.722]So the least available, what that means is
[00:18:24.862]we don't go to every country every year.
[00:18:26.901]For example, in countries with extreme droughts,
[00:18:29.511]in countries with unstable political situations,
[00:18:31.442]we might make it there in 2011,
[00:18:33.523]might not make it there in 2012,
[00:18:35.597]so this gives you the latest country-level data
[00:18:38.456]for a given metric, which allows you to
[00:18:40.717]sort of fully populate the map.
[00:18:43.802]The next page is gonna a summary page.
[00:18:46.279]This is a good place to start if instead of
[00:18:48.672]being interested in the geography,
[00:18:50.947]you're more interested in a metric.
[00:18:52.914]The recommended metrics are the
[00:18:54.694]metrics our users run the most.
[00:18:57.336]We've actually not seen a big change on them,
[00:18:59.579]and we think that's because we're actually priming them
[00:19:01.692]by telling them they're the recommended metric,
[00:19:04.178]so then people run them more.
[00:19:05.833]All Metrics will give you all the metrics
[00:19:08.076]that are available for Afghanistan,
[00:19:09.942]so you can actually scroll down here,
[00:19:12.759]and then you can click on one,
[00:19:14.306]and it will just populate that trend for you.
[00:19:16.546]So, we're gonna scroll down.
[00:19:17.835]Let's do Confidence in the Military.
[00:19:19.203]So you can see some of those country-level trends.
[00:19:21.408]You can change the country input here,
[00:19:23.311]so we're gonna look at Australia,
[00:19:25.537]and then you can look at, oh,
[00:19:28.212]I probably should relink the trend,
[00:19:29.317]and then again, this access can't be toggled, unfortunately.
[00:19:34.611]You can also export this.
[00:19:36.455]Export at the top.
[00:19:43.217]You can export them to image files as well.
[00:19:46.477]Finally, the Articles page is going to link
[00:19:48.907]to gallup.com articles, so our researchers
[00:19:51.441]are continually doing research on a variety of things,
[00:19:54.938]so if you are interested in, I think we just wrote
[00:19:58.962]an interesting article on happiness in Bhutan,
[00:20:01.486]I lied, that's not true.
[00:20:04.971]It will give you some of the things related to that country.
[00:20:07.179]This will actually take you right to the gallup.com website.
[00:20:11.280]So, that is gonna be the main features for data analysis.
[00:20:14.942]We have this Methodology link that is going to
[00:20:18.500]give you an overview of the World Poll methodology,
[00:20:21.452]and it will also give you a link to our
[00:20:23.155]World Poll methodology document.
[00:20:25.184]That document is about 84 pages long.
[00:20:27.560]It's gonna carefully describe all of our random root
[00:20:29.943]procedures that we use for selecting households.
[00:20:32.375]All of the methodology to create our income variables
[00:20:36.751]to create some of our indexes to weight the data,
[00:20:39.697]and a variety of other things.
[00:20:41.501]So, if you're really interested in the methodology,
[00:20:42.902]that's a great place to start.
[00:20:44.588]The data set details are gonna include some of the
[00:20:48.172]error measures for each country, so we're going to
[00:20:50.604]produce an N size for each country,
[00:20:53.012]a margin of error, a design effect,
[00:20:55.109]and then we're gonna tell you about any exclusions.
[00:20:58.014]The Gallup Daily tracking methodology is the same,
[00:21:00.159]so you can always click on the methodology document.
[00:21:03.369]The Gallup Healthways Well-Being Index
[00:21:05.363]is just a variety of questions we ask in the US.
[00:21:09.506]We're actually gonna go to the Gallup Brain now.
[00:21:14.112]So, the Gallup Brain is not the same interface
[00:21:17.756]as Gallup Analytics, this is actually taking you
[00:21:20.469]to a different site altogether, but if you do
[00:21:22.835]want any historical data, this would be the place to get it.
[00:21:26.186]So, I'm just gonna click on the Gallup Brain.
[00:21:29.024]I'm gonna give you a brief overview of topics and trends,
[00:21:32.408]so if you ae just interested in understanding
[00:21:35.562]tobacco and smoking, this will just link to
[00:21:41.265]what we call, it's a page on the internet
[00:21:45.573]that has, like, a brief overview of the trends,
[00:21:47.849]so that is what the topic page will bring you to.
[00:21:52.012]I'm gonna go back.
[00:21:53.495]And if you wanna do more of that true data data,
[00:21:57.624]looking at the data, I would go to the Gallup Brain,
[00:21:59.734]but unfortunately, here, you have to look by questionnaires,
[00:22:03.926]which is not as useful as the Gallup Analytics
[00:22:07.113]portal itself, but there's a lot of
[00:22:09.652]different data in here.
[00:22:10.952]It's the only place to get that historical data,
[00:22:13.447]so I'm actually gonna click on the
[00:22:17.816]Values and Belief survey, and this will bring you to
[00:22:20.040]all of the top lines of the data.
[00:22:22.871]Unfortunately, you can't do any demographic analysis,
[00:22:26.035]but Gallup is also willing to partner with you
[00:22:28.308]if you have some specific research needs
[00:22:30.901]with a question in the Gallup Brain,
[00:22:34.180]so this gives you the top line data.
[00:22:36.013]It also gives you the the N size,
[00:22:37.846]and it's also gonna show you the order
[00:22:41.172]of the questions as they appear on the survey,
[00:22:43.578]so again, these are the Gallup Poll Social Series,
[00:22:46.170]so they are VRDD.
[00:22:48.817]There are some ways to build trends.
[00:22:52.479]They're a little different.
[00:22:53.466]So, if you're interested in building a trend
[00:22:55.401]in the Gallup Brain, you're gonna click on the question item
[00:22:59.858]and that will just bring you to,
[00:23:01.354]it will tell you what questionnaire it was on,
[00:23:04.234]when that questionnaire was fielded,
[00:23:06.417]what the overall sample size was.
[00:23:08.605]Be careful, occasionally we split sample items,
[00:23:11.519]so just because the overall sample size
[00:23:12.948]might be 1,000, it could be that the question sample
[00:23:16.433]size is 500, so make sure if you're reporting
[00:23:19.289]a specific metric, that you're using that
[00:23:21.241]question's specific N size, so here,
[00:23:25.242]it will show you what the scale is.
[00:23:27.653]On this side, it will give you the overall count,
[00:23:30.206]and then, if you'd like to create a trend,
[00:23:31.656]you're going to locate questions.
[00:23:36.619]So, what's gonna happen is it's gonna bring up,
[00:23:38.612]how this tool works is it searches to match
[00:23:41.140]those scale values on the bottom with that question text,
[00:23:45.165]and you can click on all of them,
[00:23:47.774]and it will allow you to create a trend over time.
[00:23:51.860]So, currently, we're looking at it in Excel format,
[00:23:55.274]and that will go all the way back to 2007,
[00:23:58.181]so you can see the changes over time.
[00:24:00.583]You can also look at it in a chart form.
[00:24:03.366]For these large charts, you can actually play it,
[00:24:06.374]which will allow you to select different portions
[00:24:09.247]of the chart that you would like to look at.
[00:24:12.019]If you want to know what the key is, always make sure
[00:24:14.477]you're looking on the right for what those lines mean.
[00:24:18.716]I'm gonna go to the Gallup business journal.
[00:24:21.828]This is gonna link to articles that are not published
[00:24:24.303]on gallup.com, per se, but are published
[00:24:26.596]in the Gallup Business Journal.
[00:24:28.870]I'm gonna go over a couple things that are not
[00:24:30.688]in Gallup Analytics I think people have asked about.
[00:24:33.790]So, the three main sources are the only
[00:24:35.836]thing that isn't Gallup Analytics.
[00:24:37.486]Unfortunately, we don't have business-level
[00:24:39.334]or firm-level data yet, but I'm going to go back
[00:24:43.209]and show you some more of the geographies in G1K.
[00:24:48.274]So, for G1K, we have state-level data,
[00:24:53.027]but we also have MSA-level data,
[00:24:56.672]so MSA-level data, that refers to metropolitan
[00:25:01.577]statistical area, and I think Omaha counts,
[00:25:06.285]so it'll also be one, so you can look at something as small
[00:25:10.494]as a sort of city-level data, so that would be
[00:25:13.190]as granular as you can get in the tool.
[00:25:16.024]I'm gonna Clear All, so we can just look at Omaha.
[00:25:21.788]One of the things to consider is that the G1K data
[00:25:25.870]is updated every single day, however,
[00:25:28.336]we don't report any data in the tool until we have
[00:25:31.869]an overall aggregate of 300, so occasionally,
[00:25:35.072]if you're looking at the beginning of the year for, say,
[00:25:37.628]next year, you're looking in January for MSA-level data
[00:25:41.791]in 2017, we won't have gotten 300 responses yet,
[00:25:46.027]so none of the data will display in the tool.
[00:25:48.202]So, we're loading MSA data, we do two loads a year,
[00:25:52.795]and all the MSAs are also weighted to MSA targets.
[00:25:56.504]They're not weighted to overall national targets,
[00:26:00.662]that would not make sense.
[00:26:01.782]So, those are loaded twice a year,
[00:26:04.417]usually in July and then in December.
[00:26:08.465]So, if data's not showing up, that's why.
[00:26:11.118]The same way for states.
[00:26:12.698]Of course, larger states such as California
[00:26:15.204]really quickly get a large-enough sample size to report,
[00:26:17.963]smaller states or less populous states take a while
[00:26:20.435]to show up, but you can always see a national aggregate.
[00:26:24.031]You can, I'm gonna clear these time areas,
[00:26:27.925]and I'm going to just add the United States.
[00:26:32.406]I'm gonna add the aggregate.
[00:26:33.552]And so, the time period here, you can search by year,
[00:26:36.803]you can search by half-year, and that's because of that MSA,
[00:26:40.396]and so if we have enough sample on the MSA,
[00:26:43.108]you can compare the first half of the year
[00:26:46.353]to the second half of the year on an MSA,
[00:26:48.449]but that's usually for larger MSAs,
[00:26:51.072]and for the smaller ones,
[00:26:51.905]it does take the full year to populate them.
[00:26:54.275]We have about 140 MSAs that are possible to be included.
[00:26:58.944]Some years, they make the cut because we get 300,
[00:27:01.205]some years they don't, just because of,
[00:27:03.702]we're randomly dialing, and we don't exactly
[00:27:05.693]what the sample will look like.
[00:27:08.514]I'm gonna delete the MSA and I'm gonna add back
[00:27:13.609]just the aggregate, and that will enable you to look at
[00:27:18.064]a couple of different times, so you can look by year,
[00:27:20.488]you can look by half-year, you can look by quarter,
[00:27:23.642]month, week, and by day, so if we are interested,
[00:27:27.758]you can look at a specific day.
[00:27:29.745]Let's look at January 15th.
[00:27:34.268]So, you can get down to that really, really
[00:27:35.846]granular level, so 1,000 people we called,
[00:27:38.618]I'll actually scroll over.
[00:27:40.250]We do always show you the unweighted end.
[00:27:42.446]This is not the weighted end.
[00:27:43.700]It is truly what the number of people that responded was,
[00:27:47.690]so on a given day on January 15th, 2016,
[00:27:51.643]we had 7% of people strongly disagreeing
[00:27:55.108]on that they are proud of the community
[00:27:56.926]or the area where they live.
[00:27:58.461]The dash at the end, with Gallup Healthways,
[00:28:01.942]just means it's collected in partnership
[00:28:03.389]with Healthways, which is a health organization
[00:28:05.323]that sponsors some of the questions on our Daily poll.
[00:28:09.200]So, those are gonna be the time periods you can search by
[00:28:11.468]for the Dailies, and I promised I would look and see
[00:28:14.592]how far back our US Daily data goes,
[00:28:18.354]and it's gonna go back to 2008.
[00:28:23.133]That's the tool as a whole.
[00:28:24.380]Just a reminder, the World Poll data will be released
[00:28:27.666]a little bit differently, so we collect the data
[00:28:30.407]all through the year, and it's released at the end
[00:28:33.527]of every month, so it takes a while to load the data,
[00:28:36.576]so I would say, usually the first or second week
[00:28:39.102]in any given month, there'll be 20 to 30 new countries
[00:28:42.023]based on where we fielded that month,
[00:28:43.871]and what data sets have made it through the
[00:28:45.865]process of vetting and cleaning and weighting,
[00:28:47.834]so that will be updated a little bit differently,
[00:28:49.797]and we also only have the yearly data for that.
[00:28:53.181]It does take us usually until December to get those
[00:28:55.803]regional and global aggregates.
[00:28:57.740]We don't allow the tool to report
[00:28:59.855]any incomplete aggregates.
[00:29:03.617]When we first started this project,
[00:29:04.879]we let the global aggregate populate
[00:29:06.415]no matter how many countries we're in.
[00:29:08.526]We found out that, because we have a variety
[00:29:11.198]of different users, people would be making conclusions
[00:29:13.741]about the world, and we would be missing countries
[00:29:15.985]like India and China, so we wanna make sure
[00:29:18.226]that those aggregates are only populating
[00:29:19.759]when they're truly aggregates and they're truly
[00:29:21.740]representation of their global roll-up.
[00:29:26.839]That covers what I wanted to cover.
[00:29:28.786]Do you guys have any questions, either about using the tool,
[00:29:32.194]or where the data comes from, or how it's collected?
[00:29:36.809](student asking question off-mic)
[00:29:39.882]That might be a good question for David.
[00:29:43.495]There might be a few ways
[00:29:45.183]we can talk about it offline.
[00:29:51.435]What about reviewing all the different
[00:29:52.790]questions you ask in these?
[00:29:56.324]So, we are happy to send you the question list.
[00:29:58.832]There's 120 metrics, about 120 metrics,
[00:30:02.757]but that includes the demographics in about 100,
[00:30:05.548]but if you want to just see all the questions
[00:30:08.254]that are asked, I would recommend going to
[00:30:11.237]the Summary page, and then you can scroll down
[00:30:13.978]and see, I'll go to All Metrics, you can see
[00:30:17.042]all the metrics that are available for
[00:30:19.251]a given geographic area.
[00:30:32.572]For some of the weightings you mentioned
[00:30:34.402]that are used for population size, are those weightings
[00:30:38.234]reported so that we know how it's calculated?
[00:30:41.000]So, for the longest time we did.
[00:30:42.687]So, we based most of our population estimates
[00:30:45.015]on the World Bank population estimates.
[00:30:48.562]That's where we start.
[00:30:50.097]Occasionally, if we go to a country,
[00:30:52.621]so Madagascar would be a good one, there's a large
[00:30:54.654]portion of the country that, for safety reasons,
[00:30:56.898]we can't go to, so we actually do adjust
[00:30:58.980]that population size number down.
[00:31:00.925]We posted them, and we didn't have anyone
[00:31:04.159]that ever downloaded them, so we ended up
[00:31:05.446]taking them off, 'cause it was causing some
[00:31:07.066]confusion among our internal users,
[00:31:08.825]but I'm happy to provide those to you.
[00:31:10.968]We're always open about that.
[00:31:13.152]It comes in a spreadsheet form,
[00:31:14.928]and so if you're interested in population size,
[00:31:16.859]that's where that comes from.
[00:31:24.495]For the US data, you can go down to the MSA level.
[00:31:28.090]MSAs are built on counties.
[00:31:30.105]Any chance you can get down, desegregate to county level?
[00:31:33.242]No.
[00:31:34.075]Unfortunately, you can only get down to MSA level,
[00:31:37.289]and that is because of the way we're coding
[00:31:42.745]the sample when we're dialing.
[00:31:44.743]Ultimately, there might be a possible way.
[00:31:48.709]I think the problem with geography that's really hard
[00:31:50.936]is we're doing it by zip code, and so some zip codes
[00:31:55.304]are crossing county lins, some zip codes
[00:31:56.772]also do cross MSA lines, but we have just not done
[00:32:00.412]the work to get down to that level.
[00:32:04.358]If you're wondering where we get the MSAs
[00:32:06.396]and the SAV files for them,
[00:32:08.534]we're just borrowing them right from census.
[00:32:16.944]I've got access to the data drive,
[00:32:19.295]but I don't know how to open the files on the drive.
[00:32:22.855]Those are for the SAV files.
[00:32:26.698]Do I need to install any software?
[00:32:28.884]Yes.
[00:32:29.717]So, if you access to the Gallup,
[00:32:31.701]so this, I would, I'd just like to make sure
[00:32:34.364]everyone understands, this is not respondent-level data,
[00:32:36.149]so you're not gonna be able to run a linear regression
[00:32:38.470]in this program, but there are other ways to access
[00:32:41.231]the Gallup data on campus, so I believe an SAV file
[00:32:44.896]is an SPSS file, but it should be possible
[00:32:48.277]to convert to Stata.
[00:32:51.011]I've not converted to R.
[00:32:53.207]I believe how the Gallup Daily Data had been housed
[00:32:56.459]was it had just been housed on some computers.
[00:32:59.080]Yeah, you know what,
[00:32:59.913]I never tried to request access to it.
[00:33:01.763]I believe you have access to it.
[00:33:03.897]Okay, excellent.
[00:33:04.978]So, what you'll want to do is,
[00:33:08.056]do you have SPSS?
[00:33:10.190]No.
[00:33:11.023]Okay, and Technology might be able to help this,
[00:33:14.386]you can probably convert it to a
[00:33:16.309]different kind of file that you have.
[00:33:18.340]When you download that file for the Gallup World Poll,
[00:33:21.550]it's gonna be 1.5 million respondents,
[00:33:24.172]so it's gonna be a pretty large file,
[00:33:25.939]but you can do anything you want in that file.
[00:33:28.544]You can select by any country, you can select by years,
[00:33:32.593]you can, if you want to create your own weights,
[00:33:35.827]or if you wanna weight for two or three years of data,
[00:33:38.448]you can do that in the file.
[00:33:40.868]You can do any modeling you want,
[00:33:43.722]and that will truly be the respondent-level data,
[00:33:46.667]but maybe you can touch base with me and David
[00:33:49.732]and we'll figure out how to convert those files,
[00:33:51.975]'cause they are, they do in SPSS form.
[00:33:58.908]I actually think it could be possible
[00:34:01.923]to create them in a flat form, in just a CSV form
[00:34:04.845]and then you can, it tends to be
[00:34:07.230]pretty easy in SASS and Stata.
[00:34:09.602]I've been less lucky with R, getting things
[00:34:11.417]to upload from my CSV file, but that should be doable.
[00:34:15.706]Okay, thanks.
[00:34:18.825]Well, thank you so much, Elizabeth.
[00:34:19.999](audience applauds)

The screen size you are trying to search captions on is too small!

You can always jump over to MediaHub and check it out there.

Comments

0 Comments

Gallup Analytics Data Session

Description

Searchable Transcript

Comments icon comment

Related Channels

Comments