00:00:00.000 In this video, we are going to explain nominal, ordinal, interval and ratio levels of data.

00:00:05.940 Together these terms are known as the four levels of measurement in statistics. Here we

00:00:11.430 will unpack each of them in simple terms along with loads of examples so that you can approach

00:00:16.800 your analysis with confidence. Now if you are new to the oftentimes intimidating world of

00:00:21.780 academic research be sure to hit that subscribe button for loads of plain language actionable

00:00:27.720 advice covering all things research related. Also if you are working on a dissertation,

00:00:32.280 thesis or research project check out our free collection of easy-to-use chapter templates to

00:00:38.880 fast-track your research. You can find the link to those in the description. Right,

00:00:43.680 with that out of the way let us jump into it.

00:00:47.640 Okay so let us start with the basics. When you are collecting survey data or really

00:00:53.400 any kind of quantitative data you are going to land up with one or both of two types of data

00:00:59.700 categorical and numerical. These two types reflect different levels of measurement to be specific.

00:01:05.459 Let us unpack them. Categorical data as the name suggests means data that reflect characteristics

00:01:12.420 or categories no big surprise there. For example, categorical data could include variables such

00:01:18.720 as gender, hair colour, ethnicity or even coffee preference. Simply put utilising categorical data

00:01:25.740 is essentially a way of assigning number values to inherently qualitative data. For example, you

00:01:32.100 could assign number one for female number two for male and so on. Numerical data on the other hand

00:01:37.560 reflect data that are inherently numbers based and quantitative in nature. For example, age height or

00:01:44.340 weight are all things that are naturally measured as numbers. This contrasts against categorical

00:01:49.800 data which as I mentioned involves assigning numbers to qualitative characteristics. All right

00:01:56.100 so we have got categorical and numerical data as a foundation now so let us drill down further.

00:02:01.980 Within each of these main data types, there are two levels of measurement. Within categorical,

00:02:08.400 there is nominal and ordinal level data and with a numerical, there is interval and ratio level data.

00:02:16.140 But why do we need all these levels you ask? Well, the reason it is important to define types of data

00:02:23.220 is because the level of measurement for any set of data will directly impact which statistical

00:02:29.340 tests you can and cannot use on it. For example, you can only calculate the mean, the average,

00:02:35.820 for certain levels of data. If you try to run a statistical test on an unsuitable data set data

00:02:42.600 that is at the wrong level of measurement you will end up with meaningless results. So you

00:02:47.520 need to understand what you are dealing with before you can really work with the numbers.

00:02:52.440 All right so let us unpack each of the four levels of measurement along with some practical examples.

00:03:00.660 Nominal data, a categorical data type describe qualitative characteristics or groups with no

00:03:08.820 inherent numerical value, order or rank between categories. For example, gender ethnicity or

00:03:15.660 blood type, the brand of appliance someone owns, for example, a refrigerator, microwave

00:03:21.600 etcetera or someone's personal preferences for example their favourite meal or favourite colour

00:03:28.740 etcetera. In all of these examples, the data options are categorical and there is no ranking

00:03:34.740 or natural order. In other words, they all have the same intrinsic value. One colour,

00:03:40.140 for example, is not ranked above another. Simply put you can view nominal data as the

00:03:45.840 most basic level of measurement reflecting categories with no rank or order involved.

00:03:51.720 Nominal data-generating questions are often used in surveys to collect sample

00:03:57.180 demographic information such as gender and ethnicity. They are particularly useful for

00:04:03.300 comparing how responses vary among these basic dimensions. For example,

00:04:08.160 if you wanted to assess statistically significant differences in responses between men and women.

00:04:14.700 So nominal level data the most basic of the four can still be very useful depending on

00:04:21.959 your research aims. All right so now let us take a look at nominal's big brother ordinal.

00:04:29.760 Ordinal level data kick things up a notch. This type of data is the same as nominal data insofar

00:04:37.320 as it describes categories but unlike nominal data, there is also a meaningful order or rank

00:04:43.380 difference between the options. Some examples of ordinal level data include income levels,

00:04:49.560 for example, low-income, medium income or high-income. Levels of agreement,

00:04:54.960 for example, disagree, neutral or agree, levels of satisfaction, for example, poor,

00:05:02.160 average, good or excellent. As you can see in these examples all the options are still

00:05:07.980 categories but there is a natural ordering or ranking difference between the options.

00:05:13.980 You cannot numerically measure the differences between the options because they are categories

00:05:19.500 after all but you can order or logically rank them. Simply put ordinal level data reflect

00:05:26.280 categories with a natural rank order. Right, now that we have covered the two categorical

00:05:31.740 levels of measurement let us take a look at the two numerical levels of measurement.

00:05:39.420 Interval level data are numerical data in other words this level of measurement

00:05:44.400 involves naturally quantitative data, data that are by default measured in numbers. Specifically,

00:05:51.240 interval data have an order like ordinal data but in addition to this,

00:05:55.380 the spaces between measurement points are equal, unlike ordinal data. Sounds a bit fluffy and

00:06:02.160 conceptual well let us make it a little more tangible. Some examples of interval level data

00:06:08.580 include credit scores which range from 300 to 850. GMAT scores which range from 200 to 800,

00:06:16.320 the temperature in Fahrenheit, the Fahrenheit bit is important though and I will explain why

00:06:21.420 a little later. Notably in all of these examples the data points are naturally numerical they are

00:06:27.300 always expressed in numbers and the spaces between measurements are equal. For example,

00:06:32.880 the difference between a credit score of 610 and 620 is the same as the difference between 650 and

00:06:40.020 660, 10 points. However, if you look closely at all those examples you will notice that they all

00:06:46.320 have something else in common too. That something is that the zero point is arbitrary. For example,

00:06:52.740 a temperature of zero degrees Fahrenheit does not mean that there is no temperature or no heat

00:06:59.100 at all it just means that the temperature is 10 degrees less than 10. Similarly, you cannot

00:07:05.640 achieve a zero credit score or GMAT score well, let us hope not. Simply put interval is a level

00:07:12.480 of measurement that is numerical. You can measure the distance between points but that does not have

00:07:17.700 a meaningful zero point. The zero is arbitrary so interval data offer a more sophisticated level of

00:07:24.600 measurement than nominal and ordinal but it is still not perfect. So enter ratio data.

00:07:33.540 As you have probably guessed ratio is the most sophisticated level of measurement. Just like

00:07:40.200 interval level data ratio data are ordered or ranked and the numerical distance between points

00:07:46.620 is consistent and can be measured. So what makes ratio the king of measurement levels is that the

00:07:53.940 zero point reflects an absolute meaningful zero, unlike interval level data. For example, weight,

00:08:01.380 height or length you can see how a zero value is meaningful here. The temperature in Kelvin since

00:08:08.280 zero Kelvin means zero heat, a length of time, for example, seconds or minutes or hours. In all

00:08:15.480 of these examples, the zero point is absolute. Zero seconds quite literally means zero duration

00:08:21.840 similarly zero weight means weightless. In other words, it is not some arbitrary number this is

00:08:28.200 what makes ratio the most sophisticated level of measurement. With ratio data not only can you

00:08:34.740 meaningfully measure distances between data points i.e. add and subtract you can also meaningfully

00:08:42.419 multiply and divide. For example, twenty minutes is indeed twice as much time as ten minutes. You

00:08:48.900 could not do that with credit scores, interval level data as there is no such thing as a zero

00:08:54.840 credit score. So as you can imagine you can do a lot more in terms of statistical testing with

00:09:00.480 ratio level data that you can with lower levels of measurement. Okay so let us do a quick recap.

00:09:06.360 At the highest level, we have got categorical data which are more qualitative in nature and

00:09:12.780 numerical data which are quantitative by default. Within categorical, the two levels of measurement

00:09:20.040 are nominal, categories with no inherent rank and ordinal, categories with a natural order or

00:09:27.060 rank. Kicking things up a notch numerical includes the two measurement levels of interval and ratio.

00:09:33.840 Interval level data are measurable in that the spaces between measurements are equal but it does

00:09:39.700 not have a meaningful zero point whereas ratio level data the king of measurement levels does.

00:09:46.260 It is also worth mentioning that numerical level data, interval and ratio are continuous data types

00:09:53.760 as opposed to discrete. Also if you are planning to use SPSS it is worth noting that numerical data

00:09:59.940 are referred to as scale data and this covers both interval and ratio level data.