It’s time to say good riddance to super-boring, highly-repetitive, and mind-numbing-tedious grunt work. Set yourself up for success using Q’s automation features.
Here, you’ll discover how to make Q:
- Automate your data cleaning.
- Shine a spotlight on the most interesting relationships in your data.
- Whip up new variables, like filters and coded open-enders.
Meet the Automate Menu
In Q101 you discovered the right click and the Create menu. Welcome to the Automate menu!
The Automate menu can be a little full-on at times. With so many tricks and tips, it can be slightly tricky to know where to start. But follow these steps and you’ll be zooming through your analysis in no time.
TIP: You can take a shortcut by typing what you want to do in the Search features and data box at the top of your screen.
Cleaning your data (yawn)
Before you start analyzing data, you’ll often want to check if anything is looking a little odd. Are sample sizes too small? Are some questions lacking responses? Are outliers skewing your results?
To automatically run a check on the integrity of your data file go Automate > Browse Online Library > Preliminary Project Setup > Create Tables for Data Checking.
Q will then show you anything that looks ‘odd’ as new tables in the Report Tree. Data containing outliers will be shown as histograms, while any other issues will be highlighted in yellow in the relevant tables.
Now you can fix any odd data by doing things like merging together small categories (drag and drop), deleting empty categories (right click) and deleting outliers (Automate > Browse Online Library > Preliminary Project Setup > Identifying and Removing Outliers). Otherwise, just ask your data provider to fix it.
How to pinpoint interesting relationships
Imagine creating thousands of crosstabs and then removing the ones that are not interesting and then formatting them to look exactly how your company likes them – all in a matter of seconds! Here’s how:
First, create lots of crosstabs
You can do all this in one step. Go Create > Tables > Lots of Crosstabs.
Next, a box will pop up – hold down Shift and select the questions you want to place in the rows of your tables (these are usually the questions you are analyzing). Click OK. Now hold down shift and select the question you want to place in the columns of your tables (often demographic questions). Click OK.
All your crosstabs will appear in a folder called Crosstabs in the Report Tree on the left of your screen.
Second, remove crosstabs that aren’t interesting
If your crosstabs don’t contain any statistically significant relationships, it’s probably fair to say they are not interesting! Here, we show you how to delete all insignificant crosstabs with one click – so you’re left with only interesting tables.
Locate the folder that contains all your crosstabs in the Report Tree on the left of your screen (e.g., Banner Tables) and click on it to highlight.
Next, click Automate > Browse Online Library > Delete tables and plots > If not significant at the 0.05% level (95%). Tada! Your folder will now only contain crosstabs with significant relationships.
Tip: Instead of using the Automate menu you can type your query (e.g., Delete) into the Search features and data box at the top of your screen. Then select the option called If not significant at the 0.05% level (95%).
Automatically adding a top 2 box column to lots of tables
Top 2 box scores are a handy way of summarizing positive responses from Likert scales, by netting the two most positive categories in a table (e.g., Satisfied + Very Satisfied). Summarizing a scale down to one number like this makes it easier to compare results across different variables.
While you can easily drag and drop within tables, you can also automatically create a top 2 box score for all relevant questions in a couple of clicks.
Go Automate > Browse Online Library > Create New Variables > Create New Top 2 Category Variables (That is, Top 2 Boxes).
Click Yes to the first question (only show a list of questions that use a scale) and then hold down Shift and pick the questions you want a top 2 box for. Click OK and you’re done!
Creating filters and applying them in bulk
The easiest way to create a filter is to create a summary table or crosstab of the question you want to use for the filter (use the blue dropdown and brown dropdowns.
Next, click on the category you want to use for a filter (if you want several categories, hold down shift and select them all).
Now click the filter icon located on the top toolbar. It’s important to name your filter with a well-thought-out descriptive name. Click OK. Note: your filter now appears in the Filter dropdown on the bottom left of your screen.
To apply your filter to lots of tables at once, hold down Ctrl and select all the tables you want to apply the filter in the Report Tree on the left-hand side of your screen (if you haven’t created your tables yet, make sure you create them using the blue and brown dropdowns and save them to the Report Tree using the Duplicate button).
Tip: To apply a filter to all the tables in a folder in the Report Tree, just click on the folder and then select the filter from the Filter dropdown.
Coding Text data
If you code verbatim or text responses, the next steps will save you hours. If you outsource your coding, save the money and DIY. Seriously, it’s that easy in Q.
Locate the question you want to code in the blue dropdown. Now click on the blue arrow next to the blue dropdown.
This will take you to a new tab called the Variables and Questions tab. This is probably the only time you will need to be in this tab – until you are an advanced user. 😊
The row that contains your text variable will be highlighted at the top of your screen.
Next, right click on the highlighted row and Insert Variables > Code Text > New Code Frame. Now you will need to decide if you want people to be coded into only one category (i.e. single response) or more than one category (i.e., multiple responses). For this example, let’s select Pick One (single response).
You will now be in the coding screen. Take a moment to have a look around.
The first thing to do is to set up your code frames. Don’t worry, you can always add more as you go as well. Right click on the code called new code on the right-hand side of your screen and select Rename. To add another code, right click, add code and continue until you have built out your code frame.
On the left-hand side of your screen, you will see the uncoded responses. The first response will be highlighted. Just click on the code you want it in (right-hand side of the screen). That response will them be removed from the uncoded group and added to the new code. Start coding all your responses this way.
To check what responses have been coded under which codes, locate the Show responses from dropdown at the top left of your screen – this defaults to Uncoded. Click on the dropdown and select one of your new codes. The box below it (labeled with a 3 on the image above) will now only show the responses that have been coded under that group.
When you are ready, OK. Click on the Outputs tab on the bottom left of your screen. And select the coded version of your question in the blue dropdown (you can just type coded in the blue dropdown as a quick way to find it).
Tip: Text questions always have an icon showing a in the blue and brown dropdowns.
Q has already grouped any responses that are the same. You can see this by the number at the end of the uncoded response.
Now you can say good riddance to super-boring grunt work!
