Analyzing data using Means

Learn how and when to use Means, or averages, in your analysis.

Joe Razza avatar
Written by Joe Razza
Updated over a week ago

The Mean is an important calculation to have in your toolbox when analyzing survey data. Luckily, KnowledgeHound understands that and provides an automatic system for calculating and displaying this calculated field.

Isn’t the Mean also the Average?

Let’s use the following example to answer this. To find the average age of three people ages 4, 16, and 32 we use the following formula: (4+16+32)/3 = 26.

Using everyday English we say that the average age is 26yrs old. However, in statistics, we would say that 26 is the mean age.

So using the word average is perfectly fine, but KnowledgeHound uses Mean to more specifically define that we are describing the mean of the sample data you’re using in your analysis.

When is the Mean important?

Means play an important role when analyzing numeric variables, especially those with a wide distribution. For example, if you asked 100 people how much they paid for their primary house, you may almost certainly have 100 different responses over a wide range.

It wouldn’t be very helpful to know that 1% paid $145,500 for their house. It might, however, start to become more helpful to know that the average price paid is $240,000. The mean becomes even more helpful when used in comparisons. For example, if we analyze price paid based on the region we can see if a region has an impact on the average home prices people are paying.

With examples like this and others, the Mean can help make the analysis more digestible when you’re not looking for the standard distribution in your analysis.

Where do I find the Mean button?

When analyzing data you will find the Mean in the variable card listed with the other calculated fields. Calculated fields are noted with a calculator icon.

When are Means supported?

Means are supported for Numeric variable types. More specifically, KnowledgeHound will calculate the Mean for any variable where the response is a real number or if the response option has a numeric value coded to it.

Currently, Means are supported for the Base variable of your analysis. KnowledgeHound identifies the base variable with a tag saying ‘Base variable’ above it.

Now let’s look at examples of each variable type where the Mean is supported.

An example of a numeric variable would be asking respondents what their age is and asking them to enter a number. If we ask 5 people, we might get back the following answers: 19, 24, 27, 29, 34. From here, we can calculate the Mean using the five numbers.

Another example of a Numeric variable is referred to as Categorical Numeric. This variable type looks like a standard pick one, or pick many survey questions. The difference here is that behind the scenes a numeric value has been assigned to the response options. For example, if we ask respondents how likely they would be to purchase a new brand of peanut butter, the response options for this categorical question might be:

  • Definitely would buy

  • Probably would buy

  • Might or might not buy

  • Probably would not buy

  • Definitely would not buy

However, as stated above, the response options of a Categorical Numeric variable have a numeric value coded assigned to it as shown in our example below:

  • Definitely would buy (5)

  • Probably would buy (4)

  • Might or might not buy (3)

  • Probably would not buy (2)

  • Definitely would not buy (1)

With this example, we are able to use these values to calculate the Mean score.

How do we calculate the Mean?

Let’s look at our two examples in more detail to understand how KnowledgeHound calculates the Mean.

Looking at our age question, we asked 5 respondents to enter their age. The answers were 19, 24, 27, 29, 34. To calculate the Mean we add up the responses and divide by the count of responses. So in this case it is 133/5 which returns a Mean of 26.6. The same way we grew up calculating the average, because these terms, as we learned earlier, are used interchangeably.

Looking at our Categorical Numeric example we asked 10 respondents how likely they would be to purchase a new brand of peanut butter. The answers to this question came back as follows:

  • 3 selected - Definitely would buy (5)

  • 2 selected - Probably would buy (4)

  • 3 selected - Might or might not buy (3)

  • 1 selected - Probably would not buy (2)

  • 1 selected - Definitely would not buy (1)

Here we calculate a weighted average to get the mean:

((3 * 5) + (2 * 4) + (3 * 3) + (1 * 2) + (1 * 1))/10 = 35/10 = 3.5

The mean value for this response set is 3.5.

When are Means not supported?

Means are not supported for traditional categorical variables. Categorical variables account for the majority of variable types you are likely to encounter in a survey. Examples of these variables might be, but are not limited to, questions such as: What region do you live in?, Which of the following social media platforms have you used in the past 30 days?

Can I adjust the Numeric value associated with a Categorical response?

Currently, only KnowledgeHound Customer Success or Data Processing teams can alter the default Numeric values of data loaded into KnowledgeHound. This exists to ensure data integrity and to dissuade data misuse

Did this answer your question?