Joke Collection Website - Cold jokes - How to study statistics?

How to study statistics?

Other children's shoes can see the full text, the so-called experience; However, be sure to read the last paragraph.

I have studied statistics for four years (in fact, I have only studied statistics for two and a half years), no, seven years to be exact. I want to share something with you. Because I recently reached the final stage of my third grade thesis, there are an endless stream of children's shoes to ask me all kinds of data processing questions every day. I really want to answer: I don't know! Because I have to write my own paper. Therefore, it is very important to learn statistics well; And statistics is also very studious. For a person who didn't pass 140 in high school mathematics, but only took 120+ occasionally, everyone's mathematics level should not be below me.

If there are excessive words in the article, please block them yourself. I wrote this article with strong feelings. Sa, a dog!

First of all, we must establish two basic viewpoints, which will be an important source of motivation and self-efficacy for everyone to learn statistics.

First, we study psychology!

Do you know what's the use of studying psychology? I never knew. It was not until my junior year in the class that Mr. Shu taught me, that I suddenly realized. I can't study psychology, but I can design experiments! Suddenly I feel that the value is there! (Is it? = =! You know, we can experiment and design, and we will play the role of general manager, general manager and general manager. Any research begins with experimental design, whether it is qualitative design or quantitative design. In short, design is the task of the top, not the top, and it is something that only company leaders can do. Therefore, our IQ is absolutely high! Such a complicated design, if we can all figure it out, then statistics is really just a small problem. In addition, all statistical methods are based on specific experimental designs. It is the so-called party commanding the gun that needs psychology to learn psychostatistics.

Second, we don't study math!

Maybe everyone will be confused again, especially many students who have been tested by foreign departments (I don't mean to belittle other professional students here, although I always think that postgraduate psychology should get a high C! Well, after reading the updated cold jokes, hang the Southeast Branch = =! )。 Statistics is mathematics (in broad sense)? Why is it an advantage not to study math? Because, we don't need to know how we got here, we just need to care about how it doesn't exist, ah, no, how we use it. For a statistical method, we don't need to deduce its process. I haven't seen any literature that proves that ANOVA is a one-sided test or a two-sided test (do you know if ANOVA is one-sided or two-sided? ), we just need to know when it can be used, that's all!

After establishing the above two viewpoints, I fully believe that statistics is very simple (at least on the surface) in high flyers, a psychological college. You see, I never worry about the asterisk. All my research is descriptive statistics.

First, Dont Ask For Help's solution.

Don't let Sister Hongyun know that I said this, or shit will be TT. ) This is a very depressing question for many freshmen, including me. I remember a classmate in our class once said a classic saying: You won't know until you learn statistics.

Unknown is always unknown.

. Yes, it's classic! You can't figure out why statistics are like this. All statistical methods are nothing, nothing. The same data, you don't do it significantly, I do it significantly? ! You don't understand why. For the first-year students, we began to study statistics with great interest, because all the brothers and sisters told us that statistics were very important, and Sister Hongyun had a lot of homework, and so on. We also have ambitions. We almost despise statistics. What are mean and standard deviation? What do you need to learn? But suddenly one day, without any warning, we couldn't understand what Sister Hongyun was saying. Since the hypothesis test, we don't know statistics at all.

What shall we do? Of course, if you can figure out the idea of hypothesis testing, you will know what its hypothesis is. We don't need to know so much (although I am still occasionally reminiscing about the Excellence of the book Probability Theory and Mathematical Statistics), as long as we can use it. This is the best method for beginners of statistics. We just need to know under what conditions, with what method, how to put the data in SPSS, how to operate, and whether the final result has an asterisk. Isn't that what psychological research is all about?

Therefore, for a beginner in statistics, it is better not to ask for answers. For us, what we have to do is to know what is there as much as possible, not why. As for why, it is natural for statisticians to prove it, and you don't need to know whether his proof is correct or not, and how many unreasonable assumptions have been made to reach these conclusions. So from this point of view, Bayesian statistics is not difficult, just one, one, one thing! This groove is too old. Hey = =! )

Second, learn for use.

Ask the students of Yan Er a question first. Will anyone analyze it according to this? Or if I ask a question about the third and fourth grades of undergraduate course, can anyone still use Kovaldo to analyze it? I guess everyone was shocked. What's all this? Yes, what are these? ! ! I don't know if it's because I never use it or I never have a chance to use it.

You don't need to forget it.

This is the characteristic of statistics.

I have always wondered that statistics is a skill. Why can't it be as memorable as swimming or riding a bike? In fact, the statistical method we learned, that is, analysis of variance, may be the most impressive, because we have been using it. Not if you don't use other methods. Everyone is like this, so am I (even Miss Liu). A method will be forgotten as long as it is not used. If you don't watch the program for half a year, you won't know what you wrote before; It takes a lot of time to remember. Therefore, a word of advice, don't be too greedy, want to learn everything, learn what you want to use, and be sure to use it after learning. Otherwise, it is really a waste of study!

Of course, this part may be a little too radical for everyone, because you still have exams, whether undergraduate or graduate, in short, you have to deal with the past. However, from another perspective, the exam is equivalent to using it. In order to use it for exams, you must learn it!

Let me give you a hint about the exam. It seems that the final term is coming ~ (it's really coming! ) Undergraduate students have to study (a), graduate students have to study (b), and others have to wait for XX.

(A) nothing more than variance analysis and regression analysis. These two parts are the focus of all basic statistics, so most exams are compulsory; You should also know the difference and connection between variance analysis and regression analysis, which is a classic topic of psychological college (unified examination, postgraduate entrance examination and examination) and also the threshold for you to enter higher statistics from elementary statistics! Secondly, nonparametric test is actually very important, because in practical research, many times the data do not meet so many assumptions; Chi-square test is also the basis of multivariate statistical analysis in future research. Therefore, everyone still has to master it. The second semester of senior one (the first semester of statistics) may not be regression, so the analysis of variance should be more important. By the way, you need to master all the contents of hypothesis testing and parameter estimation, as well as the relevant contents of significance, and you can't forget it! By the way, the average standard deviation or something, high school should know = =! (Tucao: What if the correlation of 0.002 is significant? Is there anything I don't know? )

(2) Although the postgraduate courses will change every year, the focus will not change: the first level focuses on structural equation modeling, and the second level focuses on multilevel analysis and mainstream multivariate statistical methods. Mainstream multivariate statistics refer to multiple regression (including logistic regression), factor analysis and cluster analysis. Other non-mainstream statistical methods probably know what people are doing. Structural equation model is the key and difficult point, and it is also the development trend of psychology in recent decades. Multi-level analysis, latent category analysis, structural equation derived growth model, you can understand as much as you can, and it doesn't matter if you don't ask for answers, but you should know that statistics can already deal with so many different data, different assumptions and different design problems. For graduate students or doctoral students, publishing articles is the first priority, and good articles with good methods will be published! So it is especially important for you not to ask for a solution. But I still remind:

Don't abuse technology

Third, clarify the relationship.

For statistics, it is nothing more than a representation of the relationship between several variables. We also do research, just want to know what the relationship between some variables is. Therefore, any statistical method can be solved as long as the relationship between these variables can be clarified. Knowing these relationships clearly, we can choose the corresponding statistics.