Lesson video

Hi there.

My name is Chloe.

And I'm a geography field studies tutor.

This lesson is called Fieldwork: Analysing, Concluding, and Evaluating Air Pollution Data.

It forms part of a unit of work called Anthropocene: What Impact Have Humans Had on the World? In this lesson, we're going to be analysing our data in more detail, drawing conclusions, and reflecting on our inquiry as well, thinking about what we would change if we were to do this inquiry again.

Let's get started.

By the end of this lesson, you'll be able to analyse and reflect on your air pollution data to create meaningful conclusions and an evaluation.

There are some key words for us to think about, first of all.

Spearman's rank is a statistical test that shows the strength of any correlation between two variables.

Micro-scale inquiries are inquiries that focus on a small area, such as a street or a park.

Macro-scale inquiries are inquiries that focus on a large area, such as a city or a whole country.

This lesson is in three parts.

We're firstly going to be analysing our air pollution data, drawing conclusions from it, and then, finally, evaluating the whole study.

So let's start with that first one, analysing air pollution data.

We are now in this fourth stage of our inquiry cycle data analysis.

Geographers often start their analysis by simply describing the areas of agreement or consensus, or the areas of disagreement or conflict.

Let's have a look what Alex says here.

"The greatest area of consensus was in the public opinion of a new ULEZ scheme.

42 people, or 70% of those surveyed, thought them to be very unfavourable." Alex has done something really straightforward here, but it's very clever as well.

He's described where the majority of people had an opinion, but he's also done a small part of data manipulation as well, recognising that 42 out of 60 people is 70% of those surveyed.

He's just made our data mean a little bit more by grading it up into percentages.

Sam says, "The most favourable option was electric buses.

36 people, or 60% of those surveyed, were either favourable or very favourable about their introduction." Sam, again, has used a little bit of data manipulation there in order to create more meaning.

Once geographers have described their data, they try to explain it.

Let's look at this conversation between Lucas, Laura, and Jacob.

Lucas begins, "No-vehicle zones and a new ULEZ scheme received very negative responses." Laura says, "Maybe this is because these ideas directly affect the driver.

They could be seen as punishments." Jacob says, "Meanwhile, electric buses and cycle lanes are more about encouragement of sustainable choices, not punishment at all." He's saying that maybe that means that people would be more favourable of those ideas.

Let's check our understanding now.

Examine the stacked bar chart showing public opinions on different air pollution management strategies.

Which statement is true? A, more people were in favour of no-vehicle zones than ULEZ schemes, cycle lanes created the most even spread of opinion, or C, twice as many people voted very favourably as favourably for electric buses.

Pause the video here, so you can have a look at those statements again, and have a look at, of course, at your stacked bar chart as well very carefully, and think which one of those statements is true.

Right, let's look at those statements each in turn.

More people were in favour of no-vehicle zones than ULEZ schemes.

Let's have a look.

Well, we can see straight away that is not true.

More people were in favour of the ULEZ schemes than the no-vehicle zones.

You can see the two green colours on the ULEZ scheme is higher than it is in the no-vehicle.

Cycle lanes created the most even spread of opinion.

If we look at the cycle lanes bar, yes, there does seem to be a fairly even spread of opinion there.

Yes, some of them are greater than others, but compared to the other bars, there's no one category that really stands out as having more favorability or unfavorability than any other.

So so far, B looks like a potential option.

Let's, though, finish it off and look at C.

Twice as many people voted very favourably as favourably for electric buses.

So let's take a look there.

No, of course, that is the other way round, isn't it? We can see that twice as many people voted favourably than very favourably for electric buses.

That light green is much larger than the dark green.

So that tells us that B is our correct answer here.

Well done if you've got that correct.

Izzy is looking at the secondary data she collected on levels of asthma and air pollution in the 10 areas closest to her school.

Let's look at the columns that we've got here.

So that's all the middle column, average annual PM, or particulate matter, concentration, and that's in micrograms per metre cubed.

You can see here the data is quite varied, we got everything from 6 up to 8.

5.

Then in the next column along, we've got the percentage of the population with asthma, and again, quite a variety of different data in there as well.

The 10 areas have simply been given a letter code for the purpose of this part of your study.

From the raw data alone, Izzy is not sure if there is a relationship between the two variables.

Is it the case that the areas with the highest concentration of particulate matter also have the highest levels of asthma? She's not sure at the moment.

To check this, she decides to carry out a statistical test.

The Spearman's rank correlation coefficient test is a mathematical test that tells geographers if there's a correlation or a relationship between pairs of data.

So it's where the maths does the hard work for us.

Geographers don't have to be good at maths to use statistical tests.

What they do need to do is be able to understand the results of the test, and that's really important.

Stats tests are not something to be frightened of if you are not sure about maths.

Let's take a look at how we could carry out a Spearman's rank stats test.

The first thing we do is rank each set of data from lowest to highest.

Let's have a look at that now.

So you can see in that first column, where we're looking at particulate matter, our lowest value is six, so that gives us a ranking of one.

The next lowest is 6.

4, which gets a ranking of two.

The next lowest is 6.

7, that gets a ranking of three, and so on and so on, until we get up to 8.

5, with a ranking of 10.

We do the same thing for the column which has the percentage population with asthma, our lowest value there is 6.

1, that gets a ranking of one.

The next is 6.

9, that gets a ranking of two, and so on, up to 8.

2, with a ranking of 10.

The next thing we do is calculate the difference, or d, between the ranks, and we do this by taking rank two away from rank one.

So in our column here, we can see nine minus nine is zero, four minus three is one, and so on and so on.

Do note that we can have negative numbers here.

So three minus seven gives us negative four.

In the next step, we square the differences, so we square D, and then calculate the sum of these.

So we square our ds, and you can see that we are getting one squared is one, three squared is nine, negative four squared, of course, is 16.

So it removes on any negative values from the other column.

We sum these together, and in this case, adding all of those d squared values together gives us a total of 46.

We then use the Spearman's rank equation to find a value, n is the number of pairs of data.

So in our case here, the number of pairs of data that we've got is 10.

We've got 10 locations, 10 pairs of data.

Let's follow the equation through.

So the first thing we do is take our sum of d squared, which was 46.

We can now apply that into our equation.

I've also substituted n for our value, which is 10.

Working out the top of our equation there, you can see the six times 46 has been calculated, that brings us to 276, and then underneath that part of the equation, we've also used R 10 cubed, so that's 1000 minus 10.

This gives us 276 over 990.

Calculating that gives us 0.

279, but remember, our Spearman's rank is one minus that value.

So our Spearman's rank value is 0.

721.

We would generally go to three decimal places when we are using Spearman's rank.

The Spearman's rank value will be between negative one, which is a perfect negative correlation, and plus one, a perfect positive correlation.

So it would appear on this kind of scale between those two values.

If it's between negative 0.

7 and positive 0.

7, it is thought of having no correlation.

The correlation is too weak for us to really consider it being meaningful.

If, however, it's between negative 0.

7 and negative one, that would indicate a negative correlation.

And if it's between positive 0.

7 and positive one, it would be a positive correlation.

As Izzy points out, "Our Spearman's rank value is 0.

721.

This indicates a positive correlation between air pollution and levels of asthma, but it's a relatively weak correlation," 'cause it's quite close to that 0.

7 value, which means that we can only really take it as being a possibility of there being a correlation between the two.

It doesn't show a really strong indication of a relationship between air pollution and the level of asthma.

Let's check our understanding now.

True or false, geographers have to be good at maths to do statistical tests? Think about everything I've just said.

Is it true or false? Pause the video, and then come back to me.

Right, let's see what you got.

Well done if you recognise it's false.

Now, tell me why.

Yes, geographers do not need to be good at maths, but they do need to understand what the result of their statistical tests mean.

Let's practise now.

Look at your stacked bar chart.

Describe and explain what you see in the graph.

Then carry out a Spearman's rank correlation coefficient test on a selection of 10 pairs of air pollution data and asthma level data that you have taken from secondary data sources.

Describe what your Spearman's rank correlation coefficient result tells you about any correlation between air pollution and levels of asthma.

Now, naturally, you're going to want to pause the video here.

You've got quite a lot of work to do.

Take your time over that stats test, and check it with a friend as well to see if they get the same result, then come back to me, and we'll talk about what those results mean.

So your first task was to look at your stacked bar chart and to describe and explain what you see in the graph.

Your answer may include something like this, "The greatest area of consensus was in the public opinion of a new ULEZ scheme.

42 people, or 70% of those surveyed, thought them to be very unfavourable.

This could be because a ULEZ directly affects the driver, and it could be seen as a punishment." You are then asked to carry out a Spearman's rank test and to comment on what the results tell you.

Here's something your answer may have included, "Our Spearman's rank value of 0.

721 indicates a positive correlation between air pollution and levels of asthma, but only a weak correlation." Now, of course, it depends on what your Spearman's rank value was, as to what you are saying in this section.

Hopefully, you managed to find a result which had some meaning.

Now, let's move on to the second part of this lesson, where we're gonna be thinking about our conclusions.

So concluding air pollution data.

Here we are in the fifth stage of our inquiry cycle, the conclusion.

Geographers begin their conclusion by reviewing the main points of their analysis.

They need to decide which of their observations are most important and have the strongest evidence to support them.

Geographers can then answer their inquiry question.

Let's remind ourselves of our inquiry question in this case.

Should people be more concerned about air pollution in our local area? Andeep is reviewing his analysis and the strength of his evidence, and he makes some notes.

Strong evidence includes people tend to be both worried and apathetic about air pollution.

He got that from his word cloud.

Highest air pollution data was found nearest the traffic lights.

He got that from his choropleth map.

People like management strategies that encourage sustainable choices.

That was something that he managed to find in his questionnaire.

The weaker evidence as well is important.

Andeep concluded that high levels of poor health conditions, such as asthma, are not always found in areas with high levels of air pollution.

He can conclude that from his Spearman's rank.

By drawing on the data that produces the strongest evidence, Andeep is now ready to write his conclusion and answer his inquiry question.

To form well-rounded and meaningful conclusions, geographers often have to try to combine the results of different data sets together.

Aisha says, "Though people are worried about air pollution, they don't want vehicles to be removed from the roads, even though we found that it was actually the road junctions that had the highest levels of air pollution." In this statement, Aisha is referring to three different aspects of her data.

Can you spot all three aspects in her statement? Let's take a look at those now.

First of all, she's got data from her word cloud, "Though people are worried about air pollution," so she's picked out that piece of data, but she's compared that against what their choices were with regards to management strategies.

She said, "They don't want vehicles to be removed from the road." So this interesting contradiction, that they're worried about air pollution, but they don't want to stop driving on the roads.

She said that this was interesting because, actually, it was the road junctions that had the highest levels of air pollution.

She got that from her choropleth data.

So all of those three things together have actually formed a really well-rounded conclusion, because she's able to combine the ideas across different data sets.

Why should geographers try to link elements of their data sets together? Is it A, to show where there are areas of weakness in the investigation, B, to make their data sets larger, or C, to make their conclusions more well-rounded and meaningful? Pause the video and have a think, and then come back to me.

So what answer did you get? Hopefully, it was C.

Yes, geographers try to link elements of their data together to make their conclusions more well-rounded and more meaningful.

The conclusion then addresses any hypotheses made at the start of the inquiry.

Let's remind ourselves what Jun said his hypothesis was.

He said, "I hypothesised that air pollution will be highest around road junctions, so people will be keen to reduce air pollution here, as we live in a built-up area." And we can see we've got the choropleth map here to help us make that decision.

Jun's hypothesis can be partially accepted.

His data does show that air pollution was most concentrated around the road junctions, but one cannot say if everybody wants to actively reduce air pollution.

In fact, his questionnaire showed quite varied results in that regard.

Sofia has this hypothesis, she says, "I hypothesise that people might be angry about the amount of air pollution near the roads, but I don't think that they will be willing to change their behaviour to stop it." Sofia's hypothesis can be accepted.

Some people did speak about being angry about air pollution, however, they also disliked some strategies to reduce air pollution, especially those that directly affected their habits.

So she's right in having the idea that people might be angry about something, but they won't always want to change their behaviour.

The acceptance of Sofia's hypothesis highlights a common aspect of fieldwork inquiries that are based on opinions and human behaviour.

There are not always geographically logical outcomes.

As Sofia says, "We can see from our map that vehicle emissions are the most likely cause of air pollution in our area, and people are very worried about this, yet they're not in favour of no-vehicle zones.

It doesn't make sense." Should Sofia be worried about writing a conclusion based on contradictions, where it seems that one message is given in one question, but a totally different message is given in another? Part of being a geographer doing fieldwork is accepting that there will be lots of opinions on local issues that don't always make sense, or which may be very different to our own.

These contradictions are extremely helpful to geographers who help make decisions about the way we live.

It helps them to be creative and come up with solutions to problems in a way that will be engaging and successful.

Let's now check our understanding of that.

Complete the sentences with the missing words.

You should pause the video here, so you can have a scan through the paragraph below, and then come back to me when you've got three words that can fit into those gaps.

Let's check what words you found.

Geography fieldwork that investigates public opinions and human behaviour may not always produce logical outcomes.

These conclusions are still valuable or valid, as they help geographers to engage with the public in their decision-making.

Well done if you've got those.

Now, our second practise task of the lesson, write a conclusion in relation to your own data analysis.

State whether any hypotheses you made are accepted, partially accepted, or rejected.

Do pause the video here to go back over your data, so you can form really well-rounded, meaningful conclusions.

So what did you get in your conclusion? It depends on what your data, of course, says.

But your answers should include a statement that answers your inquiry question, a summary of the strongest evidence, and a statement that says whether your hypotheses are accepted, partially accepted, or rejected, and importantly, why you are doing that with your hypothesis.

We now move on to the third and final part of this lesson, evaluating a geographical inquiry.

Here we are, in the final stage of our inquiry cycle.

In every fieldwork inquiry, there are things that go well and as expected, and other things that do not go to plan.

Alex and Sam are reflecting on their fieldwork inquiry.

Alex says, "I'm really pleased we managed to get 60 people to answer our questionnaire.

It is a good-sized sample on which to base our conclusions." I completely agree with Alex here.

60 people is a great number of questionnaires being answered.

If it was fewer than that, maybe we wouldn't be able to draw as good conclusions as we have.

Sometimes external factors that are out of the geographer's control can have an influence on the results.

Factors, such as weather conditions, cannot be managed easily.

As Sam says, "On the day we put out the particulate matter capture card, it was quite windy, so they might have captured particulate matter from outside our local area." And of course, that would've have affected the results.

Let's check our understanding of that.

Which of the following are external factors that could have had an impact on the amount of particular matter captured on the cards? And remember, external factors are things that you cannot control.

Have a read through the options there, pause the video, and then come back to me with the right answer.

Okay, let's take a look at the answers.

Well done if you got those three, the wind direction, the time of day or year, and rainfall, those are all external factors that would've had an impact on your cards.

Now, all of the others would possibly have had an impact as well.

The size of the capture card could make a difference, the height of the card, the number of cards that were put out, and the type of cardboard used might have affected how the card was able to capture different bits of data.

But of course, we could control all of those things.

Air temperature probably wouldn't have had any impact at all on the amount of particulate matter that the cards would've picked up.

But wind directions certainly would've done.

It could have been blowing from an area which had a lot of pollution initially, or had a source of pollution, and that could have actually affected how much was recorded on the card.

The time of day or year would've had an impact, potentially, there's time of day particularly, it could be during rush hour, there's lots of vehicles on the road, that means that it could be that there's a lot more particulate matter in the air than perhaps on a weekend, where there would be less.

And of course, rainfall.

If the card were to get wet, perhaps it could be washed off some of the particulate matter, so that would've had an impact as well.

Well done if you got those.

Laura says, "It's a shame that we couldn't directly link our capture card data to our secondary data on asthma levels." As Lucas said, "The two data sets are at completely different scales.

The capture cards are part of a micro-scale inquiry.

Health data tends to be at a macro-scale inquiry level.

This means they cannot be fairly compared." It's a really strong observation from Lucas there.

The capture cards were simply put out around your local area, around your school, whereas health data is collected on a much wider scale across the country.

It's very difficult to compare data that you've collected locally with data that is collected nationally.

Geographers think a lot about how scale influences their inquiry findings.

They ask questions like this, "Would my results be the same at a larger or smaller scale?" "Is the scale of my inquiry too large to show local details?" "Are all aspects of my data comparable according to scale?" "Is the scale of my inquiry too small to see bigger trends?" So true or false, air pollution data taken at a micro-scale is a reasonable indication of the air pollution at a macro-scale? Have a think about what we've just discussed.

Is that true or false? Pause the video here, and then come back to me.

Yes, of course, that is false.

But why is that the case? Yes, the data is not comparable because it is of different scales.

It would be wrong to assume that all areas of a micro-scale inquiry have the same air pollution levels of that of a macro-scale inquiry.

Let's finish off this lesson by our final practise task.

Write an evaluation based on your air pollution fieldwork inquiry.

Include the following points, a discussion about something that went well, a discussion about how a way external factors may have influenced your data, and a discussion about how the scale of the investigation might have influenced the results.

Think carefully about those three points before you put pen to paper.

Pause the video, and then come back to me.

Let's take a look at what you have written.

Your answer may include something like this, "Something that went well is that we managed to get 60 people to answer our questionnaire.

It is a good-sized sample on which to base our conclusions.

On the day we put out the particulate matter capture cards, it was quite windy, so they might have captured particulate matter from outside our local area.

This is an external factor that may have affected the results.

The air pollution data from the capture cards was at a completely different scale to the secondary data we collected on asthma levels.

This means the two data sets cannot be compared." Let's conclude the lesson by summarising our learning.

Geographers often use statistical tests, like the Spearman's rank correlation coefficient, to see the nature and strength of any correlation between two data sets.

They also combine the findings from multiple data sets together to form well-rounded and meaningful conclusions.

Geographers are mindful of the influence that external factors, such as weather conditions and the scale of studies, can have on their data.

Well done for all of your efforts in that lesson, particularly if you gave that Spearman rank stats test a good go.

It's not easy to approach these things the first time, but the more you do them, the more confident you will be in using stats tests as a way of analysing data in more detail.