• Tidak ada hasil yang ditemukan

BUSINESSSTATISTICS DEMYSTIFIED - MEC

N/A
N/A
Protected

Academic year: 2023

Membagikan "BUSINESSSTATISTICS DEMYSTIFIED - MEC"

Copied!
399
0
0

Teks penuh

You may use the work for your own non-commercial and personal use; any other use of the work is strictly prohibited. Your right to use the work may be terminated if you do not comply with these terms.

Nonparametric Statistics 298

Our approach is to focus on understanding statistics and how to use them to support business decisions. You will learn how to use statistics in real business situations, and how to prepare top quality statistical reports for your assignments.

HOW TO GET THE MOST OUT OF THIS BOOK

We wrote this book so that you can apply statistics in a practical way.

SIDEBARS FOR EASY LEARNING

Instead, we have only partial information—too often inaccurate information—about what is happening in the real world. Both try to tell us what we are likely to find in the real world or what is likely to happen if we do.

Fig. I-1. Business statistics, mathematics, probability, models, and the real world.
Fig. I-1. Business statistics, mathematics, probability, models, and the real world.

CHAPTER

It is also important to be able to determine which general features of the world will help us make our decision. Which of the following is not an advantage of understanding how to lie with statistics. a).

WHAT IS MEASUREMENT?

Event. The particular instance of the particular act of measurement, usually identified by the combination of the subject and the time at which the measurement is taken. Each value usually describes an aspect or feature of the subject on the occasion of the measurement.

LEVELS OF MEASUREMENT

In algebra, different types of numbers are defined in terms of the different possible values ​​contained within them. Statisticians do not classify different types of measurements in terms of what types of numbers (or non-numeric symbols) are used to record the results.

Table 2-1 Automobile inventory.
Table 2-1 Automobile inventory.

CLEANING YOUR DATA

In order to make decisions, we need to know the true value of the information statistics provide. If we are using the computer, we may want to use the full names of the colors of the cars in the lot.

TWO WAYS OF BEING WRONG: VALIDITY AND RELIABILITY

Now we go into the hot air balloon and take a picture of all the golf balls, right above our heads. If all the numbers are close, we know that our technique for measuring Judy's height is reliable, but how do we know if it is valid.

WHAT IS A POPULATION?

The easiest way to make a good study bad is to not specify the population of interest clearly and precisely. Sampling. The process of selecting the individuals from the population that make up our sample.

WHAT IS A SAMPLE?

In this context, random means that each member of the population has an equal chance of being selected as part of the sample. This is a form of convenience sampling in which the subjects determine whether or not they are part of the sample.

SUMMARIZING DATA: WHEN IS A NUMBER A STATISTIC?

As you can imagine, a descriptive statistic is one that describes a general feature of the data. All guarantees for that statistical technique come with the proper use of inferential statistics.

WHAT IS A STATISTICAL TECHNIQUE?

Which of the following is the end product of the procedures used to calculate a statistical measure. a). By the final sections of the chapter, we will be ready to make our first real statistics.

LIKELIHOODS AND ODDS: THE MYSTERIES OF CHANCE

One way we do this is with the concept of 'odds'. If the chance is small, we say the odds are against it. Suppose we want to express the number of black sheep, n, in terms of its ratio to the total number of sheep in our flock, N.

PROPORTIONS AND PERCENTAGES

Also, key to probability theory is the distinction between the ratio between the subjects of interest and the rest of the rest, as calculated by subtraction in probability. Note that, mathematically, if pi is the proportion of subjects of interest to the total, then 1p is the ratio of subjects that are not of interest.

COUNTING POSSIBLE OUTCOMES: THE RULE OF

INSUFFICIENT REASON FOR CLASSICAL PROBABILITY

In statistical theory, the proportion of the entire population is a theoretical model of probability (at least according to some probability theories). The rule of insufficient reason does not mean that we have good reason to believe that all basic outcomes (such as the appearance of one side of a head, or a particular sheep being gathered) are equally likely to occur.

REPLICATION AND THE FREQUENCY APPROACH

In the context of an imaginary simple experiment, it is the act of assigning a single value to each variable in such a way that the probability of different values ​​is changed by the very act of sampling. We can think of the results as if we hit any black sheep with the lasso and any white sheep with the lasso.

COMMON SENSE AND SUBJECTIVE LIKELIHOODS

The name Bayesian probability is a bit misleading, because Bayes' Law can be applied to any type of probability. There are times when none of the assumptions hold and any notion of probability is questionable.

STATISTICAL INDEPENDENCE: CONDITIONAL AND UNCONDITIONAL LIKELIHOODS

Even here we have to be careful, as units coming off the assembly line one by one can tend to cluster together in the bin. It's a Jack of Spades, which is bad for us and we lose the hand. Then we say: 'It's the jack of spades.

WHAT IS A PROBABILITY DISTRIBUTION?

The truncation of the curves (c) The position along the x-axis (d) The kurtosis of the curves. The truncation of the curves (c) The position along the x-axis (d) The kurtosis of the curves.

Table 3-1 Heights of Judy and friends.
Table 3-1 Heights of Judy and friends.

WHERE TO GET DATA: THEFT VERSUS HONEST TOIL

And the main driver of cost is how we collect the data. Our data collection plan tells us where and how to get the data into the form.

HOW TO GET DATA: METHODS

The content of the questions will help you determine the population, sample size, and where to get the data. The question we must ask is which of the following methods is the least expensive acceptable source for our data and statistics.

OTHER PEOPLE’S DATA

With this data, as collected, you meet the requirements of the assumptions of the statistical procedures you want to use. Then define which tables, graphs and charts will display the data in the statistical report.

PREPARING A BUDGET

Use the same headings as in the summary and write a detailed, step-by-step procedure for how you will prepare the report. This topic goes beyond this book, but we can recommend another text from Sid Kemp's Project Management Demystified series.

GETTING THE PLAN APPROVED

That way, if you're in a rush at the end, you can dump the data, run the stats, dump the results, and the report will be ready to go. If so, you can plan things that will make the study easy to repeat and show the value of reducing the cost of repeating the study in the future.

FREE STATISTICS

A big advantage of using other people's statistical results is that sometimes detailed data cannot be provided for security or privacy reasons. If the provider could not calculate certain statistics we need, there is no way to calculate those statistics without detailed data.

NOT-SO-FREE STATISTICS

The only question that usually arises is whether it is better to use someone else's data or collect it yourself. If the vendor has a good reputation and our statistician informs us that the methods used by the vendor to collect data are suitable for our intended statistical analyses, then obtaining the vendor's data is probably preferable to collecting our own.

FREE DATA

A very clever seismologist realized that the information he needed was collected 24 hours a day in the urban areas of Japan. All he had to do was to obtain the post-earthquake data from the area where the earthquake occurred.

NOT-SO-FREE DATA

Security cameras in stores provided moving images from which the seismologist could gather detailed data on Earth's movements. By notifying hundreds of stores in the city in advance and ensuring that the tapes were preserved after an earthquake, the seismologist could obtain ample data the next time the area was hit by a serious earthquake.

DATA YOU ALREADY HAVE

Furthermore, collecting our own data means that we can collect exactly the data we need using the methods most suitable for our statistical analyses. We evaluated the data and statistics the site provided and worked with it to prioritize the business questions we could answer.

COLLECTING DATA FROM YOUR OWN OPERATION

The other author of this text had a business customer come to him and say: ''We get all this data about how our customers use our site. By addressing these concerns, we can create an environment where the business generates the data it needs as part of ongoing operations.

CLEANING UP OUR DATA

The most important issue in conducting a survey is defining our population and ensuring a representative sample, as we discussed in Chapter 2. Good question design is the most difficult challenge in a survey, and we address that in Chapter 15.

REDUNDANCY, VALIDATION, AND CALIBRATION

The most reliable form of data entry is probably an electromechanical device, where a piece of mechanical equipment, such as a bimetallic strip thermometer or infrared detector, measures the temperature or length of products coming off a conveyor belt, or some other physical attribute. We can use computer input devices such as scanners to enter survey data in a similar way.

STORAGE

If they were to fill out paper forms, there would be physical evidence of this, such as the handwriting pattern. So we would have to devise other means, such as a hidden timestamp on each survey, showing when the survey was conducted and how long it took, perhaps linked to the telephone system used for the survey, to identify such sources of error. tracks.

DATABASE DESIGN

There must be a single, reliably measured variable that measures the proportion of the whole for each part. If the order of the values ​​is important, another type of chart (perhaps a bar chart) should be used.

Fig. 7-1. Pie chart: breeds of sheep in our flock.
Fig. 7-1. Pie chart: breeds of sheep in our flock.

ORIENTATION OF THE BAR CHART

When the length of the bar indicates something other than count or quantity, a horizontal bar chart should be considered. Another important use of the horizontal bar chart is in the side-by-side bar chart.

Fig. 7-2. Vertical bar chart: breeds of sheep in our flock.
Fig. 7-2. Vertical bar chart: breeds of sheep in our flock.

PARETO CHARTS

Sauron will allow only the dead to rise to the highest positions of power, leadership and recognition, becoming members of the nine horsemen. While statistics cannot determine causality, it may be that this disparity of opportunity is one of the aspects of Sauron's realm that makes it legitimately considered evil.

HISTOGRAMS, AREA CHARTS, AND FREQUENCY POLYGONS

A good photo has a majority of pixels in the middle of the range, not too dark, not too light. If the ups and downs of the distribution are not clear in the histogram, change the number of intervals and try again.

Fig. 7-5. Brightness histogram.
Fig. 7-5. Brightness histogram.

SEGMENTED AND STACKED BAR CHARTS

Likewise, too few large intervals pile up the data in the middle of the graph and also hide the differences. Since men and women are sensible divisions of the group's people (in terms of the gender variable), the segmented bar chart makes sense.

MULTIPLE BAR CHARTS

Note that a vertically segmented bar chart is often called a stacked bar chart. To compare a stacked bar chart and a multiple bar chart of the same data, compare Fig. The basic line graph shows the relative frequencies of values ​​for a single variable, as does the bar graph.

Figure 7-8 could be used to illustrate the 1:10:100 rule, as follows. In Fig. 7-8, we see the expenses of three projects, broken out by project stage.
Figure 7-8 could be used to illustrate the 1:10:100 rule, as follows. In Fig. 7-8, we see the expenses of three projects, broken out by project stage.

SINGLE-LINE GRAPHS

For rank, interval, and ratio variables, the line in a line graph shows the change in score over the sequence of values ​​of the variable. Two of the most basic changes are the choice to show the points, or to show only the line.

MULTI-LINE GRAPHS

We can show the points if presenting specific figures is important, or to illustrate our data set. We smooth the curves for two reasons: to show a prediction; or to show whether our data is normally curved, distorted or truncated.

AREA GRAPHS

In short, the relationship between weight and height is evident from the shape of the point cloud in the scatter diagram. The regression line is a description of the relationship between the variables and is discussed in Chapter 12.

Fig. 7-12. Scatter plot of weight and height of Judy’s friends.
Fig. 7-12. Scatter plot of weight and height of Judy’s friends.

MIXING CHART TYPES

Often the best chart does not fall into any standard category, but may be a mix of different standard chart types or inspired by aspects of the standard types we have seen. Here we will look at mixing chart types, using 3-dimensional graphs, and creating statistical maps.

VOYAGE TO THE THIRD DIMENSION: 3-D CHARTS

See how it's easier to get your point across when women's and men's heights are displayed side-by-side on the same 3-D chart. Once everyone is on the same page, discussing different solutions will be much easier than it would otherwise be.

Figure 7-14 has a vertical axis (or Z-axis) of count, an X-axis (heading to the right) of various heights of people, and a Y-axis (heading to the left and forward) of gender
Figure 7-14 has a vertical axis (or Z-axis) of count, an X-axis (heading to the right) of various heights of people, and a Y-axis (heading to the left and forward) of gender

MAPS

Adding bias through 3-D. Volumes give a different impression of variance or change than more representative 1-dimensional lines. For example, states with a Republican governor may be in one color and states with Democratic governors in another color.).

DOING IT RIGHT

A simple vertical bar chart To compare values ​​with each other, especially when comparing quantities or value. Parallel or Mirror Bar Chart To compare two sets of values ​​where the value names are the same but the populations are different.

Table 7.1 Chart types.
Table 7.1 Chart types.

ERRORS TO AVOID

Second, anything to the right of the equal sign is just shorthand for the rules for. When we use as a measure of the amount of information in our sample, we call it degrees of freedom.

Table 7.2 Graphical elements representing statistics.
Table 7.2 Graphical elements representing statistics.

RATIOS, ALSO KNOWN AS PROPORTIONS

Finally, we need to understand the critical importance of ratio to the statistical process called estimation. For this reason, descriptions of the sampling distribution will also describe (via the estimate) the population distribution.

THE RANGE: MINIMUM AND MAXIMUM VALUES

If the sample is large enough, the sampling distribution will resemble the population distribution. The range statistic is a poor measure of the variability of a sample because it depends on only two values.

Table 8-1 Heights of Judy and female friends (sorted).
Table 8-1 Heights of Judy and female friends (sorted).

TYPICALITY: MEANS, MEDIANS, AND MODES

Gambar

Fig. I-1. Business statistics, mathematics, probability, models, and the real world.
Fig. 3-1. Breed of merino.
Table 3-1 Heights of Judy and friends.
Fig. 3-4. Male and female height distribution (bar graph).
+7

Referensi

Dokumen terkait

Namun, kita juga harus ingat bahwa kita juga adalah orang-orang yang sudah mendapat anugerah Tuhan – diampuni dan ditebus, hati kita telah dibaharui oleh kasih/Injil Kristus, Allah Roh