So you have completed a survey and you want to be able to assess how probable it is that a respondent will say yes or no to one of your variables. With a little bit of SPSS training, this is a very simple thing to do. 

For example, you have conducted research on whether females and males, with and without jobs, are likely to be pet owners. The fact that these variables are dichotomous (each is simply a yes or no answer) makes it a little easier to check the odds of who is most likely to be a pet owner while using SPSS. When your variables have a wider range of possible responses, it gets a little more complex, but that’s something for another blog post. Let’s just start simply for now.

First, enter your data. Create three variables on the variable view tab called Pet Owner, Employed and Gender and fill in your responses on the Data View tab. Remember, you do not need to put ‘yes’ and ‘no’ responses into the Data View spreadsheet. By using the Values tab on the Variables View you can team numerical data with descriptions, so you could team ‘male’ with 1 and ‘female’ with 0, for example.

Now you need to perform a binary logistic regression analysis. Click on ‘Analyse’ in the toolbar and then ‘Regression’ and then ‘Binary Logistic’. Put ‘Pet Owners’ into the dependent variable space and then Employed and Gender into the Covariables space, then click ‘Ok’.

You will now be presented with a list of tables. The most important ones you need to look at are the Classification Table in Block 0, which will tell you how many people, unemployed and either female or male, are actually pet owners. Then look at the Classification Table in Block 1, which will show you what numbers the data predicted. If you have a good model to make predictions with, the numbers in Block 1 will be similar to or exactly those in Block 0, but if the predictions are way out, you may need to reconsider what data you use and how you deploy it.

Click here for further details of Acuity Training's SPSS training courses.

So here is a question that gets asked quite often in our SPSS training courses - what exactly is the null hypothesis and how does it apply to may statistics.

Each time that we want to run a statistical test we need to formulate two hypotheses, and we then use the statistical test in SPSS to test between those two hypotheses. The null is the simplest explanation or hypothesis and the one that the researcher is trying to disprove . The alternative is usually the hypothesis that we want to test either for to prove or disprove depending on the circumstances. 

Choosing an example to illustrate this. If we took a sample of 10 people at random from a larger population that we assumed that had equal numbers of men and women in it and found that 6 were female and 4 male, then the null hypothesis would be that there are equal numbers of men and women, whereas the alternative would be that there are more women than men in the population.

Having formulated our hypotheses we then run a test to test to test the hypotheses. Given that SPSS will know the distribution for the results of the test we can then calculate a p-value (significance) value for this data. This tells us how unlikely our result would be  if the null hypothesis is true. If the p-value is small then we reject our null hypothesis.

The subject of p-values and their interpretation will be the subject of a further blog post later in the week. So check in later in the week to see the next in our series of questions that come up frequently during our SPSS training courses.

Click here for more details of Acuity Training's introductory SPSS training courses.

So you have completed a survey and you want to be able to assess how probable it is that a respondent will say yes or no to one of your variables. With a little bit of SPSS training, this is not a hard function to perform.

Let’s say, as an example, you have conducted research on whether females and males, with and without jobs, are likely to be pet owners. The fact that these variables are dichotomous (each is simply a yes or no answer) makes it a little easier to check the odds of who is most likely to be a pet owner while using SPSS. When your variables have a wider range of possible responses, it gets a little more complex, but that’s something for another blog post. Let’s just start simply for now.

First, enter your data. Create three variables on the variable view tab called Pet Owner, Employed and Gender and fill in your responses on the Data View tab. Remember, you do not need to put ‘yes’ and ‘no’ responses into the Data View spreadsheet. By using the Values tab on the Variables View you can team numerical data with descriptions, so you could team ‘male’ with 1 and ‘female’ with 0, for example.

Now you need to perform a binary logistic regression analysis. Click on ‘Analyse’ in the toolbar and then ‘Regression’ and then ‘Binary Logistic’. Put ‘Pet Owners’ into the dependent variable space and then Employed and Gender into the Covariables space, then click ‘Ok’.

You will now be presented with a list of tables. The most important ones you need to look at are the Classification Table in Block 0, which will tell you how many people, unemployed and either female or male, are actually pet owners. Then look at the Classification Table in Block 1, which will show you what numbers the data predicted. If you have a good model to make predictions with, the numbers in Block 1 will be similar to or exactly those in Block 0, but if the predictions are way out, you may need to reconsider what data you use and how you deploy it.

One of the first things people notice during our SPSS training courses is the similarity between SPSS and spreadsheets like Excel. Indeed, the data entry page you are presented with is much like a basic spreadsheet, but there are a few differences that need to be grasped before you can get to grips with the software.

In SPSS the horizontal rows are dedicated to each item of data (for instance each respondent to a survey), while the vertical columns are to be used for each variable (such as their height, weight, or IQ - whatever data you have collected).

While this might seem simple enough, it is important you do not just jump in transferring data to the spreadsheet. First, you need to learn a little about preparing the vertical columns.

If you do not name your columns, SPSS will choose a default name like VAR00001. To change the names to ones that are more meaningful to you, click on the ‘Variable View’ tab in the bottom left of the screen. SPSS will not allow spaces or punctuation to be used in the name column here, so if you need to separate words, use an underscore.

While you’re there, make sure your columns are prepared to format your data properly. Click on type and select between string and numeric. String should be used if you are going to be entering letters or words and you do not want SPSS to recognise the data as a number.

You might also need to change the width too - that’s the number of characters SPSS will allow you to enter for each entrant in that variable. Decimals refers to how many decimal points you would like to include after your numerical data, the Label is a space to briefly describe your variable (with spaces allowed this time!), Values allows you to team up numerical and descriptive terms (for instance where 1=male and 2=female) and Measure allows you to tell the programme whether your data is nominal, ordinal or scale.

The Column and Align fields are more to do with the presentation of the data, to make it easy to read. Just like in an Excel spreadsheet, you can decide how wide the columns should be and whether the text should be aligned to the left, centre or right.

Make sure you have prepared your columns in the Variable View before you start entering statistics in the Data View, otherwise you could soon find yourself swimming in unlabeled data!

There are more tips from our SPSS training courses coming soon.

Looking for easier, more efficient ways to use SPSS? As a complex and comprehensive program, even those who are used to the software may be missing out on some of the tips and tricks that can make analysing data with SPSS a piece of cake. In fact, it’s often some of the more basic ideas that catch people out, even after they have been on an SPSS training course. Check out our top five SPSS tips below for some quick and practical hints.

1) Save your master copy

It’s a really good idea to back up your original data so you have a master copy to return to should you wish to take your analysis down a different path. Leaving a paper trail as you go along, by saving as a new file each time you make a significant or permanent change to your dataset, is another top tip, as this allows you to revert back to different data transformations and analyses as and when you need to.

2) Start simple

Don’t jump in at the deep end with complicated mathematical analyses without finding out what simpler information you can obtain from your data set. Start with a descriptive analysis to provide a context to more complex procedures, and use the information you glean from this to guide you when investigating further patterns.

3) Keep your software up to date

To get the most out of SPSS, make sure you are keeping your software up to date. This does not necessarily mean buying the latest version of SPSS. The manufacturer releases updates and patches on the program frequently and downloading these updates ensures your SPSS software remains effective. To download patches, click on Help, then Check for Updates.

4) Take advantage of the temporary command

The temporary command does exactly what is says on the tin - it performs temporary data transformation which is then reversed after the first procedure you run.

This can be a valuable tool, particularly if you are analysing large datasets that could take quite a bit of time to run a syntax on. Instead, it is possible to perform a transformation on a subset of the data as a test, before running the same analysis on all of your statistics.

5)  Keep an eye on your output files

SPSS allows you to have several output files open at once, but it will only transfer data to the one that is “active”. Make sure you are aware which file is active so you can keep an eye on where your analyses are ending up.