Cross Tabs with SPSS Tutorial (SPSS Tutorial Video #7)

welcome to data demystified i'm jeff gallick and this is my series of tutorial videos and how to use spss to work with data in this video i'm going to show you how to use crosstabs to help you understand the relationship between two categorical variables as always we'll be using the youtube viewing habits survey that i created and you can find both a link to the data file and a video tutorial of this data below crosstabs are an incredibly powerful tool for comparing the relationship between two categorical variables they're not really appropriate for continuous variables so we're not going to use those here i want to jump right into it and run a crosstab that compares two things one is modality one which is what people report as their most frequent modality in watching youtube if we click into this little ellipse we see that there's a few options computer tablet smartphone tv other and not in a moment i'm actually just going to restrict this to only be those first four options honestly that's just to keep things simple you can of course include everything if you want the other variable i'm going to look at is gender which is down here so gender is coded as 1 equals male 2 equals female 3 equals other again just to keep this simple i'm only going to look at male and female you can of course look at all the data i'm just trying to keep this as an easy example for you so to restrict those i'm going to use the select case tool which i talked about in a different video but i'll do it quickly here so data select cases if and i've already written the conditional but i'll show you what it is if i go into here i'm saying exclude gender so don't allow gender to be three so it can only be one or two and i'm saying only include modality one responses that are less than five which will include our other and nones and again this is just to keep things simple you can include everything if you want so i'm gonna click continue and i'm gonna click ok and again if you don't recall how to do the select case tool there's another video i can point to down below and so now i'm going to go ahead and run my crosstab crosstab is under analyze descriptives crosstabs a couple of things to point out here there's a rows and a columns feature you can have more than one variable in each of those and it'll just create multiple sets of crosstabs again to keep things simple i'm just going to have one crosstab it also really doesn't matter what you put into rows versus columns it's completely symmetrical it's just going to pivot the table one way or another so you can put whatever you want in either of those two boxes so long as there's something in each of them so what i'm going to do is i'm going to take modality 1 and put it into my rows and i'm going to find gender by hitting g and i'm going to put that into columns i will ask for a few options under statistics i always make sure to ask for the chi square test this is actually going to tell us if there's any difference in the composition of those two variables as they relate to one another there are plenty of reasons to check some of these other options but they're pretty unusual so i'm not going to be including them in this video if that's something you want to know more about please leave a comment below and i'll make sure to create a video for that in any case i click continue under cells i'm going to ask for a few things i'm going to ask for the expected count which is going to tell me what would be expected if there were no association between these two variables i'm going to ask for the row and column percentages because i actually like to make sure i can look at this from both directions and then critically i'm going to ask for the adjusted standardized residuals this is going to give me a sense if the value in a particular box in my crosstab which you'll see in a second deviates from what we'd expected to if there were no association the rest i can actually leave as it is so i click continue and i click ok and i get my result so first as always we see the number of cases there's only 991 because i've excluded those folks who either said other for gender or other or none for modality one that's just our nine people but we still have almost everybody in our data set i'm going to skip over the crosstab itself and i'm going to look down here at this chi-square test the most relevant thing to look at here is that first row the pearson chi-square test what this is is a test of association between these two variables in this particular example we can think of it as is the ratio of male to female respondents the same regardless of whether they used a tablet or smartphone or tv or computer to watch youtube or those different and if i have a significant result here if that value is less than 0.05 i can reject the null hypothesis that there is no association and instead it sounds like there might be a difference in ratio of male to female across these different modalities to understand what that is though we have to take a look at our table so we're going to start really simple to unpack this table first we're going to look at the outside values the totals sometimes called the marginal values what these are are the representations of each of those values ignoring the other variable so what do i mean by that for the variable modality there were 342 people who said they used a computer that represents 34.5 of all of our responses within this data set another 70 people said they used tablets which represents 7.1 percent another 436 said they use smartphones which is 44 another 143 use tvs which is 14.4 and here's our total 991 people like we saw above which is 100 looking at the other variable male versus female gender we see that there were 360 male respondents or 36.3 percent of all of our responses and there were 631 female responses or 63.7 percent this tells us nothing yet about whether there's a dependency or relationship between our two variables but it gives us a description of what these variables are comprised of so now we get to the interesting piece which is is there a dependency meaning is there a difference in the ratio of male to female responses as a function of whether they use computers tablets smartphones or tvs to do that we look at each of these cells separately here's one cell what this cell is telling us is that if we did not have an association we would expect based on the ratio of males and the ratio of people using computers to see 124.2 observations in this box right here but we see more we see 148. in fact we can say that we see significantly more because our adjusted residual is above the conventional cutoff of 2.0 that's the cutoff we tend to use to claim statistical significance it's actually just a rounding of 1.96 which might sound familiar to you if you know your normal distribution more precisely the way to think about this is that overall they were 34 and a half percent of people who use computer but when we look within the category of mail 41.1 percent of them used computers that's more than our 34 and a half percent and so because this adjusted residual is above 2 we say that there's significantly more males using computers than we would have expected if there was no dependency between computers and modality and it turns out that the opposite is true for females that negative residual says that that 30.7 percent is significantly less than this 34.5 percent looking at tablets by the way we see there's really nothing going on those adjusted residuals are either below 2 or above negative 2 and so there's really no deviation meaning 8.3 is not statistically differentiable from 7.1 percent the overall value and six point three percent again is not differentiable from the overall value of seven point one percent for smartphones we again see a pretty big difference males had thirty six point one percent of them using smartphones which is considerably less than our forty four percent and we see that our adjusted residual is less than negative two and for females 48 and a half percent of them used smartphones which is more than our 44 and again significantly so because we see that our adjusted residual is above 2.0 and finally for tv there's absolutely no difference it's actually strikingly no difference we saw that 14.4 of all the people that we sampled use television to watch youtube and that's actually the exact ratio for males and females as well so really nothing going on there if i had to summarize this result i'd say that males are more likely to use computers and females are more likely to use smartphones so that's how we do a crosstab and at this point i think it'd be a good idea for you to pause the video and try this for yourself particularly what i'd like you to look at is not modality relative to gender but rather modality relative to political identification so i'll pop over here we have a variable called paul id which is political identification and that is categorized as republican democrat independent libertarian and other and we'll just leave them all in place as is so why don't you go ahead and pause the video give that a try and i'll show you what it looks like when you're done okay so let's give this a try we're going to go back up to analyze descriptives crosstabs i'm going to take gender out of here and i'm going to replace it with political id i'll leave all my other options intact but just to remind you under statistics i keep chi-squared under cells i keep the expected row and column percentages and the adjusted standardized residual and i click ok there's a lot more going on here partly because there are many more values and levels to political id than there were for gender but if we look at our chi-square table down here there's actually no dependency the pearson chi-squared is not statistically significant it is above 0.05 so we cannot reject the null hypothesis that there's no dependency between these two meaning that the use of different modalities does not significantly differ as a function of political ideology but let's take a look at this anyway so this right column right here is exactly the same as we saw before it's just the total values for our different modalities right down here is the percentages representing each of these political ids so we had 24.9 republican 41 percent democrat 27 independent 3.4 libertarian and three and a half percent other and if we note inside the boxes inside these values none of those deviate from what we would expect not significantly so anyway so for instance if there were 34 and a half percent people who used computer none of these values here even though they look like they might differ do not significantly differ from that 34 and a half percent the same is true for this 7.1 percent going across or this 44 going across or this 14.4 going across it's worth pointing out we can do the exact same thing looking the other direction for example there are 24.9 percent republicans well that ratio is maintained whether it's tv which is similar or smartphone which is similar or tablets which is similar or computers which is similar the point is there is no dependency in this particular case the ratio of political ideology does not differ as a function of what type of device people use to watch youtube crosstabs is a very powerful tool for comparing categorical variables and i hope you understand it better that's it for this video i hope you found this useful and if you have any questions please comment below and i'll be sure to reply as quickly as i can aside from these tutorials i'm on a mission to equip everyone with the information they need to thrive in our data rich world if you'd like to learn not just the mechanics of analysis which these video tutorials focus on but also learn the intuition behind the analysis you're performing i strongly suggest you check out the other intuition-focused videos on this channel where i take the jargon out of statistics and data science and help you build a deep intuitive understanding behind all the analysis that you're performing i'll put a link below to a playlist of the videos that focus on just this finally please take a moment to like the video subscribe to this channel and click that little bell icon so you don't miss out on any new content that i put out thanks for watching

Channel: Data Demystified

Views: 1,077

Rating: 5 out of 5

Keywords: spss, spss 26, spss tutorial, spss for beginners, spss tutorial for data analysis, spss data, spss real data, learn spss, learn spss 26, introduction to spss, intro to spss youtube, spss tutorial youtube, data demystified, data spss, spss walkthrough, spss how to, how to use spss, spss introduction, advanced spss, spss advanced, spss tricks, interpret spss, cross tab, cross tabs spss, cross tabs, how to interpret crosstabs in spss, crosstabs, how to do crosstab in spss

Id: 82nGWJCv6zA

Length: 10min 37sec (637 seconds)

Published: Sun Oct 18 2020