Вы находитесь на странице: 1из 4

ETW1000

Business and Economic Statistics


Lecture 9.2

In this lecture, we are going to spend some time talking about your assignment presentation in coming
weeks. In particular, we will discuss some general advice about giving presentations, recap some of
the key statistical concepts for your presentation and talk about peer assessment.
__________________________________________________________________________________

1. Multiple Regression: The Drivers of Food Shortages
__________________________________________________________________________________

To learn in more detail about multiple regression we are going to use the data from last semester’s
assignment. Last semester’s assignment was similar in spirit to that for this semester.

Students were asked to undertake a small investigation of some data from households in Africa, to
learn something about the social and economic conditions these households face. The data set
comprised the results of a household survey undertaken in 2014 of more than 1,700 poor, rural
households in Kenya, Uganda and Tanzania. These households are all in rural villages and small farms,
relying on their small piece of land for most of their food, income and daily needs.

There were three topics:
• Household crop production
• Food shortages
• Children and schooling

Let us focus on the second topic; food shortages. Our key focus is to try to figure out, in a regression
analysis, what drives food shortages. Our dependent variable is “Number of months in past year
experienced food shortages”. We have a wide range of potential independent variables. Let us run a
regression with all of them and interpret the results.

• What is this thing called a regression?
𝑦 = 𝛽$ + 𝛽& 𝑥& + 𝛽( 𝑥( + ⋯ + 𝛽* 𝑥* + 𝑒
o The coefficients are chosen to provide a ‘line’ of best fit through the data.
𝑦 = 𝛽$ + 𝛽& 𝑥& + 𝛽( 𝑥( + ⋯ + 𝛽* 𝑥*
o The error is zero on average.
• What do the coefficients tell us?
o They have a quasi-experimental interpretation; “what would happen if variable 𝑥&
increased by one unit and all other variables stayed the same?”
𝑦, = 𝛽$ + 𝛽& 𝑥& + 𝛽( 𝑥( + ⋯ + 𝛽* 𝑥*
𝑦- = 𝛽$ + 𝛽& (𝑥& + 1) + 𝛽( 𝑥( + ⋯ + 𝛽* 𝑥*
𝑦- − 𝑦, = 𝛽&
o The constant is a little different. It is the predicted value of 𝑦 when 𝑥& = 𝑥( = ⋯ =
𝑥* = 0.
• Statistical significance of the coefficients (standard error, t-statistic, p-value and confidence
intervals).
• Is the model any good?
o This is a tough question:
§ Good relative to what?
§ Good for what purpose?

1

o R Square (𝑅 ( ) and the regression standard error provide some insight into how well
the model fits the data. That is, how close the predictions are to the actual
observations.
§ 𝑅 ( always rises (or at least never falls) when you add more variables.
§ The standard error needs to be interpreted in the context of the mean of the
dependent variable.

SUMMARY OUTPUT

Regression Statistics
Multiple R 0.314353788
R Square 0.098818304
Adjusted R Square 0.090852644
Standard Error 3.401600253
Observations 1713

ANOVA
df SS MS F Significance F
Regression 15 2153.145743 143.5430496 12.40553842 8.55703E-30
Residual 1697 19635.79063 11.57088428
Total 1712 21788.93637

Coefficients Standard Error t Stat P-value Lower 95% Upper 95%


Intercept 3.7298 0.4153 8.9820 0.0000 2.9154 4.5443
Number of people in household 0.0466 0.1414 0.3298 0.7416 -0.2306 0.3239
Household head completed primary school 0.1607 0.1949 0.8244 0.4098 -0.2216 0.5431
Household head completed secondary school -0.3884 0.3207 -1.2109 0.2261 -1.0175 0.2407
Kenya Region B 1.0072 0.4082 2.4676 0.0137 0.2066 1.8077
Tanzania Region A 1.6310 0.3626 4.4983 0.0000 0.9198 2.3421
Tanzania Region B 2.0284 0.3118 6.5046 0.0000 1.4168 2.6401
Uganda Region A 1.8561 0.4222 4.3959 0.0000 1.0279 2.6842
Uganda Region B -1.0808 0.4918 -2.1974 0.0281 -2.0455 -0.1161
Number of working-age adults in household 0.0620 0.1564 0.3964 0.6919 -0.2447 0.3687
Number of children in household 0.0899 0.1475 0.6091 0.5426 -0.1995 0.3792
Size of Land (Hectares) -0.1842 0.0511 -3.6024 0.0003 -0.2845 -0.0839
Number of people aged over 60 in household -0.0336 0.1696 -0.1984 0.8427 -0.3663 0.2990
House has a new metal roof -0.5503 0.2999 -1.8353 0.0666 -1.1385 0.0378
House has an old (leaking) metal roof -0.2930 0.1993 -1.4705 0.1416 -0.6838 0.0978
Crop production (tonnes per year) -0.2284 0.0581 -3.9286 0.0001 -0.3424 -0.1143

__________________________________________________________________________________

2. Some Advice About Presentations
__________________________________________________________________________________

• Think about the context of the presentation:
o What is the purpose of this presentation?
o What is being asked of me?
o What am I trying to achieve in this presentation?
o What does the audience want, what can I get out of it?
o Who is going to be in the audience?
• Nerves?
o Practice, practice, practice.
o Why get nervous; “The consequences of a bad presentation are very minor”.
• What makes a good presentation?
o A bit of introspection is useful here:
§ In the presentations you have seen what makes them good or not so good?

2

o For me: an interesting problem, a creative approach which is clearly and logically
outlined and results that matter in practice.
• Some practical advice
o Keep the slides simple and clear
§ Limit the number of words.
§ Try to use pictures, diagrams, figures, tables where possible to get your
point across (as they say: “A picture says a thousand words”) but keep these
simple too.
o Ask yourself for every slide:
§ What point am I making here?
§ How can I make the point in the clearest/simplest way?
§ What does it contribute to the overall theme?
o Stick to the time limit!!!
• Remember, you probably won’t give a faultless presentation first time around. Keep
thinking and working on it.
__________________________________________________________________________________

3. Marking and Peer Assessment
__________________________________________________________________________________

• In preparation for what we are expecting, and to get you thinking more about the
assignment, we are going to have a go at doing some evaluations in the lecture.
• There are two presentations that we’ll have a go at:

Presentation 1: Children and schooling in rural Africa
Presentation 9: Investigation on the total production of crops in the household

3





3 = Fail (0-4 marks)
5 = Pass (5 marks)
6 = Credit (6 marks)
7 = Distinction (7 marks)
9 = High Distinction (8-10 marks)

Scan the following QR code:


OR

Key in the following URL: www.bit.ly/ETW1000projectevaluation

Вам также может понравиться