Over representation of any group in movies or TV can give people a skewed version of reality in the real world. I prefer to view the world accurately. Here's a lgbtq article about that very thing.
https://news.gallup.com/poll/259571/americans-greatly-overestimate-gay-population.aspx
Boy where to begin with several of your posts.
1st. We as a humans, have no means of measuring how many people are not what you would call heterosexual or have a gender identity that doesn't match the the gender at birth or the gender listed on birth reports. Period. As the only means of measuring this information is by self identification and has only been done by sampling. Unlike places in the world that try and make an accurate (key word try) count of the population of a country, there currently isn't a system of going door to door and asking each person their sexuality or gender identity.
For example, as someone who has worked for the US Census Bureau for almost a decade, I am versed in how data even the door to door attempt to count the entire population of the US isn't fully accurate. And most of the census's work isn't a door to door accounting but a statistical sampling based on its ten year census for various other issues (like the jobs report, the housing survey, ect). For the first time in 2021 the US census actually included in the Household Plus Survey a question about sexual identification and gender identity, with the addendum that it only counted those 18 and older in the household (as you must speak to an adult and you can't assume that an adult would know their Childs sexual orientation or gender identity).
This report is compiled yearly, but data is taken and compiled every two weeks. The survey takes forever and is very detailed (actually asking people to have their last years bills available to go over). that survey of a home took me on average about 35 minutes to do. If someone actually supplied the full amount of data requested to go over, it could easily take hours. I had one go for over 3 and a 1/2 once...The first year it came to 85 % of the adults sampled fully as heterosexual with the gender identity that matched that of their birth or birth record. The sample size of the survey varies between your home representing 450 other homes to as much as 4000. The ones I worked on averaged between your home representing 2000 to 2500 other homes. its one of the largest statistical samplings done that include sexuality and gender identity in the US and done in person. You remain part of the sample minimum for two years, with rotations occurring on 4 year and 10 years.
And I can personally attest that in my area I knew gay people who did not identify as gay in the sample. Same as I knew people of hispanic, latin or Spanish origin that wouldn't report that information accurately to the government because they simply didn't believe that information wouldn't become public. It becomes public record after 75 years.
Now there have been more targeted statistic sampling of age groups that have pointed to recently having more younger people self reporting having a gender identity or sexual orientation that doesn't match their birth gender or being straight. Only time will tell if those numbers are indicative of how society is more open to differences that more and more people are willing to either be open, or even simply identify as something different than the norm.
2. When talking about representation are you talking about trying to match the global population or the population of your country of origin or even the production company of origin. That makes a sizable difference. Since I am from the US, the US is what I have knowledge of. And since a significant amount of media for the world is made by US entertainment companies (even if filmed in locations outside the US, that one that I use a lot). Now if you are comparing representation to the world, tv and film are vastly under showing almost all non white non heterosexual characters to a staggering degree.
Now most people watch a tiny segment of the media that is created or aired in the US. With tens of thousands of shows, one person or family has no means to accurate measure the demographics of the overall industry. But there are business that do just that. And in the US with US production companies gender identity, non heterosexuality. race, and gender still do not equal the most accurate information that the US Census can provide.
And of course they also break down things like the demographics of just the cast, versus the demographics of every character shown. And in neither case does the overall marketplace give enough representation to minorities. Though its has been getting better, especially of some groups. They also breakdown things like scripted content versus reality based content, and also the combined overall. Now one addendum, there have been yearly spikes for certain demographics, where the numbers for a minority group has exceeded the most current known statistics, but they have never maintained that level, historically.
Now as individuals our specific viewing habits very likely don't match the overall market place For some you will see lower than average results, for others you might see higher than average numbers. Some viewers might only consider primary cast, others will see the larger tapestry of all characters shown on a show. As an example in the 70's almost no one on tv was identified as gay, lesbian, trans or bi directly. They did use coded language to indicate that it was likely. But when characters started actually being directly referred to by others are stating for themselves, it was almost always a guest character. When it took to the cast, it first hit supporting characters, and not the leads. Eventually it finally started showing some leads. Which is even to this day fairly rare.
3. Yes skewed representation can have impacts on people and society. And it can be both good and bad. Seeing an more positive view of your own situation can give your encouragement to make your world better. The opposite side is seeing a world more positive for your situation can put you in serious risk if you assume your society is going to behave in the manor of the programming your watching.
In much the same way that media presents a view of humanity where people are vastly more in shape and vastly more attractive then the general population, this can inspire a person to make better life choices, but it also absolutely can cause series body image problems that can lead to someone taking their own life.
These issues have been part of tv and films since their very inception, and continue and in some ways (body image especially) have gotten worse.
And one of the other significantly under represented groups is the elderly on tv and film. Dramatically under the percentage they make of the population in the US.
The Gallup report is correct that in general people due tend to thank their are larger number of minorities in the US. I have no idea on the breakdown on what the belief is of those demographics of a world wide level.