Many years ago in college, I took a statistical analysis course and learned how statistics can be used to say one thing and interpreted another way. Through that course I learned that one must clearly document how the statistic was derived, what related assumptions were made, what the sample size was, what types of people were targeted to provide answers as part of the statistical group that was sampled. , how the results were obtained. should be interpreted and the intention with which the created statistic was to be used.
Taking these thoughts further, I’d like to start with a couple of “statistical terms” that are used within this field of analysis to help me establish my starting point for this article.
Appeal to ignorance: A logical fallacy: taking the absence of evidence as evidence of absence. If something is not known to be false, assume it to be true; or if something is not known to be true, assume it to be false. For example, if I have no reason to think that someone in Tajikistan wishes me well, that is not evidence that anyone in Tajikistan wishes me well.
Average: An ambiguous term. It often denotes the arithmetic mean, but can also denote the median, the geometric mean, a weighted mean, which is derived from including external factors and weighting those factors to obtain a result, among other things. Be careful if any report you read cites an average without making it clear which the average is quoted.
Bias: A measurement procedure is said to be biased if, on average, it gives an answer that differs from the truth. The bias is the expected difference between the measure and the truth. For example, if you step on the scale with your clothes on, that biases the actual weight measurement to be higher than your actual weight.
Convenient example: A sample drawn for convenience; it is not a probability sample. For example, I could sample opinions simply by asking my 10 closest neighbors. That would be a convenience sample, and unlikely to be representative of a true target audience of significant size that could give me results that could actually be used as part of a serious survey. Convenience sample surveys should always be avoided if one is interested in publishing the truth with a reasonable degree of certainty.
You get my take on what can be done to manipulate the stats if you’re trying to make a point that you want to make. Quite often, stats are shared in this way with others to try to generate favorable behavior in certain directions, usually to the benefit of the creator of the stat.
Now, I’m not saying that the statistics are generated with the deliberate intent to mislead. But even if the recipient of the statistical information misinterprets them, you have failed in your job of using the statistics to impart accurate and usable information that could be used by the consumer who ingests the statistics.
Let me give you a real-life example of how both ambiguity and possible intent to deceive can creep into the creation of statistics that I often find represented in places like presentations, infographics, videos, etc.
I recently saw a stat about sharing content on Facebook. The title of the graph read: “Images are shared more on Facebook” and showed a graph that represented that about 80% or more of the total actions on Facebook are made to share images. the range of 5% shared content.
A person who thinks superficially after seeing this statistic might say to themselves, “Wow, I better focus on posting images on Facebook instead of videos.” What the statistic, which was basically a simple bar chart, didn’t say was: who and how many people were sampled to generate this statistic; The reason this level of image sharing occurs is because people want to see more images than videos or it’s possible that:
- So few people know how to create and upload videos and that videos take a lot of work to create so few get uploaded?
- Or is it that everyone has a mobile capable of taking photos these days and that many of the 1.8 billion Facebook users are uploading pictures of their family, their pets, their vacations or where they had dinner the night before?
As a professional business person and internet marketer, on the other hand, I know that Facebook loves to see you upload videos, and if you upload one, it will “reach” a lot of people without me having to “boost” the post. so they can see it out there; whereas with image ads, I often have to send people images and “boost your post” for a fee paid to Facebook to get them spread as wide as my videos. A recent example: A text and image ad I ran recently without boosting the post reached only 22 people. A video I recently uploaded to Facebook that same week reached 1,368 people without the boost being posted.
So my conclusion from looking at the shareable “bar chart” stat I saw the other day is that someone who read it without understanding what it meant could make bad business decisions because of a stat like that and start generating ads based on in images while completely ignoring the creation of videos. Both types of content have their uses as advertisements for a business, but keep in mind that getting your images in front of people will cost you more ad dollars than a video uploaded to reach equivalent amounts of people.
So what was the person pushing that statistical picture out into the world trying to insinuate? What was his goal? And how should I have read and used that newly shared information for my benefit? The clarity around the post just wasn’t there…
In closing, I’d like to give the following advice to those who publish statistics: If you’re a marketer trying to drive home a thought or idea for others to understand through a stat, please:
- Understand how people would react when they see the stat
- Be of integrity in presenting statistics to others as you market to them.
- Add verbiage to explain what the stat means – the gist of what it says when there isn’t enough clarity within the stat itself.
