My aim is to undermine the theory that happy workers are more creative than miserable ones, writes   Lucy Kellaway

For the first four days of last week I kept a diary of my private thoughts and feelings while at work.

The last time I did anything of this kind I was 13; the result then was banal and embarrassing and I hid it from everyone. The latest effort is also banal and embarrassing but this time I am reproducing it below.

My aim is to undermine the newly popular management theory that says happy workers are more creative and productive than miserable ones. This theory is supposedly proved by an article in the May edition of the Harvard Business Reviewin which 238 professionals were asked to keep daily diaries of their inner working lives.

Two professors pored over 12,000 entries and compared them with the quality of the work that the people had done.

I have just conducted the same experiment on myself and an analysis of my results points to a different conclusion. First, the raw data.

Monday: Feeling quite perky by recent standards. Started writing my agony aunt column with enjoyment. Felt smug. Had long gossipy lunch with colleague. Lost momentum in the afternoon. Finished column in a panic. Realised I'd forgotten to pick up my youngest son from school. Cycled home feeling less smug.

Tuesday: Slept badly. Slouched into work feeling tired, flat and thoroughly out of sorts. Had conversation with tiresome colleague and decided I loathed him. Hardly had any e-mails. Felt unpopular. Wrote some of the Martin Lukes column with distaste. Read the result and thought it laboured and unfunny. Decided I hated my job and contemplated retirement.

Later at home did some maths with my daughter. Felt oddly cheered up by simultaneous equations.

Wednesday: Lovely ride to work in the sun feeling carefree. Looked over previous day's work and thought it not too bad. Finished column with satisfaction. Chatted to people. Had lunch with another colleague. We sat in the sun and moaned, which was nice. E-mailed cheerfully. Frittered time. Day ends a little blank, but fine.

Thursday: Slightly hungover, almost no sleep. Headache, depressed. Had pointless and bad-tempered e-mail exchange with someone I like. Sat down to write this column. Too wretched to want to talk to anyone. Worked with dogged sense of purpose, though felt the result feeble beyond redemption.

Readers may be thinking two things. First, that my experience is less significant than that of 238 professionals. To this, I'd like to refer them to the pappy quality of their diary entries. Here is an example: "My boss's boss came by, which was nice. He brought bottled water!" Mine reads like Virginia Woolf by comparison.

A more serious objection is that my entries seem to support the theory that I'm trying to disprove: on my miserable days - Tuesday and Thursday - my work was bad too. The truth is more complicated.

For years I have been monitoring my vacillating moods and the effect they have on my work. When I am glum I invariably judge my work to be bad. When I'm cheery I tend to think my work fine. However, I am a hopeless judge of what I do.

Indeed, the columns that people have liked best have often been the ones written in a mood close to despair, while many of the ones I've liked have fallen entirely flat.

There are three reasons for this. When tired and unhappy I have no energy for chatting. I am grimly focused on what I'm doing. I'm much more prepared to take creative risks too - if life seems already bad, what is there to lose? And the fact that I judge myself so harshly when down makes me try much harder.

If I am right, there are surely interesting management implications. Managers should concentrate on making us as miserable as they can. Actually my diary doesn't quite show that.

There are many kinds of bad feelings and some are more productive than others. Mild depression may be good for work. Severe depression tends not to be good for anything at all. Bad moods generated not by existential angst but by cretinous managers and daft management initiatives are also unproductive. They simply make one think: why bother?

My diary seems to imply that managers should hire professionals who are by nature somewhat neurotic and depressive. But even this isn't so, because of something academics call "emotional contagion". A happy mood spills over to others and a miserable one does too.

Reading my entries on Tuesday and Thursday I can conclude that I wasn't helping the inner working life of those around me at all.

There is only one incident in my diary that has any constructive message for managers. Look at what I did on Tuesday. Notice how the maths cheered me up. This strikes a chord with the HBR survey, which found that workers were happier when they knew precisely what they were supposed to be doing and were allowed to do it.

Ambiguity is bad. Clarity is good. This explains why the equations were such bliss. I knew what I was meant to do. I had the tools to do it. And I got the answers entirely, objectively, indisputably right.